Blunted cardiovascular and catecholamine stress reactivity in women with bulimia nervosa

Psychiatry Research 80 Ž1998. 13]27

Blunted cardiovascular and catecholamine stress reactivity in women with bulimia nervosa Jeannie H. Koo-Loeb a , Cort Pedersen b , Susan S. Girdler a,b,U a

Department of Psychology, CB 7175, Uni¨ ersity of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA Department of Psychiatry, CB 7175, Uni¨ ersity of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA

b

Received 8 August 1997; received in revised form 23 April 1998; accepted 30 April 1998

Abstract Cardiovascular and catecholamine responses to mental stressors were investigated in women with bulimia nervosa ŽBN. and in healthy control subjects. Fifteen women with BN and 15 control subjects completed psychosocial questionnaires before laboratory testing, where they were exposed to an interpersonally based speech stressor and a serial math task. Blood pressure, heart rate, epinephrine, norepinephrine and, via impedance cardiography, systolic time intervals, cardiac output and total peripheral resistance were measured at rest and during stress. Results indicated that BN was associated with blunted sympathetic activation in response to mental stress, indicated by increased pre-ejection period responses and blunted systolic blood pressure, heart rate and epinephrine responses. In contrast, women with BN had elevated cortisol levels when compared with control women. In addition, despite equivalent performance between groups, bulimic women reported feeling significantly more confused, frustrated, inadequate and dissatisfied with their performance during tasks. Psychosocial questionnaires also indicated that women with BN perceived more stress, had worse coping skills, lower self-esteem and sense of mastery, reported less social support, had worse mood, had greater anxiety and were more depressed when compared with control women. These results are interpreted as reflecting physiological and psychological profiles indicative of distress vs. active effort coping in BN. Q 1998 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Eating disorders; Cortisol; Blood pressure; Heart rate; Pre-ejection period

1. Introduction It is well accepted that stress, particularly negative, emotional stress, plays an important role in U

Corresponding author, Department of Psychology, University of N.C. School of Medicine, Campus Box a7175, Chapel Hill, NC 27599-7175, USA. Tel.: q1 919 9662544; fax: q1 919 966-0708.

the etiology andror exacerbation of many psychiatric disorders, including bulimia nervosa ŽBN.. For example, women with BN often report perceiving more psychological stress as compared with non-eating disordered individuals ŽCrowther and Chernyk, 1986., and they are characterized by a lack of successful coping skills, perhaps further compounding the stress perceived wsee Cattanach and Rodin Ž1988. for reviewx. Additionally,

0165-1781r98r$19.00 Q 1998 Elsevier Science Ireland Ltd. All rights reserved. PII S0165-1781Ž98.00057-2

14

J.H. Koo-Loeb et al. r Psychiatry Research 80 (1998) 13]27

women with BN may exhibit increased levels of cortisol ŽHudson et al., 1982; Copeland et al., 1988; Mortola et al., 1989., a hormone indicative of physiological stress. Studies also suggest that stress and negative affect may precede binges and the binge]purge cycle ŽBarrios and Pennebaker, 1982; Abraham and Joseph, 1987; Lingswiler et al., 1989. and may even play a role in the maintenance of this disorder ŽCattanach and Rodin, 1988.. Despite this evidence for a role of stress in BN, the empirical evidence has yet to indicate the exact nature of the stress]illness relationship. This has been due, in part, to the lack of consistent strategies for examining the neurobiological processes of stress in psychiatric patients ŽBriere, 1989.. Thus, one purpose of the present study was to employ well-validated, psychophysiological techniques to investigate cardiovascular as well as stress-relevant neuroendocrine measures in women with BN relative to healthy control women. Evidence for the dysregulation of the stress response in BN may help to clarify the impact that greater life stressors combined with poor coping skills wsee Cattanach and Rodin Ž1988. for reviewx has on the clinical presentation of BN. It has been proposed ŽKling et al., 1989; Chrousos, 1992; Chrousos and Gold, 1992. that frequent or prolonged exposure to stressors in susceptible individuals may result in dysregulation of the generalized stress response ŽGSR., when the usual counter-regulatory forces designed to maintain homeostasis fail. In addition to the peripheral changes elicited during the GSR, which are designed to promote an adaptive redirection of energy, the hypothalamus and locus ceruleusrNE containing neurons project to other brain systems, including the mesocorticalrmesolimbic and amygdalarhippocampal systems, which play profound roles in setting the levels of arousal and emotional tone in the organism. Dysregulation of the stress response would increase vulnerability to subsequent stressors since either hyper- or hypo-activation of the stress response would decrease the range of arousal for optimum sense of well-being and performance ŽChrousos and Gold, 1992.. Indeed, there has been recent identifica-

tion of several major illnesses that occur as a result of, or are associated with, dysregulation of the stress response, including melancholic depression ŽBriere, 1989; Kling et al., 1989; Chrousos, 1995.. Although a number of studies have examined stress-related neuroendocrine measures in BN, with most, though not all, indicating reduced norepinephrine levels wGeorge et al. Ž1990. and see Pirke Ž1996. for reviewx but elevated cortisol ŽHudson et al., 1982; Copeland et al., 1988; Mortola et al., 1989., these studies have relied on basal levels only. Only two studies ŽCattanach et al., 1988; Pirke et al., 1992b. have investigated physiological responses to interpersonal or psychological stress in BN. Cattanach et al. Ž1988. found that those who scored high on an eating disorder questionnaire vs. those who scored low rated an increase in desire to binge following standardized laboratory stressors which included Interpersonal Conflict, Stroop, Speech and Social Interaction conditions. Though Cattanach et al. Ž1988. found no differences in cardiovascular reactivity to stressors, their groups were based on questionnaire assessment and were not verified, via structured interview, to meet clinical criteria for any eating disorder. Pirke et al. Ž1992b. investigated differences in norepinephrine ŽNE. levels in hospitalized women with BN vs. control women during a control Žno stress . session and during a mental stress session. In that study, women with BN had significantly increased basal levels of cortisol but decreased NE during the control session. While performing the mental stressor Že.g. noise during a computer-based intelligence test., women with BN showed a significantly blunted response for both cortisol and NE compared with control women. These investigators were led to question whether the blunted norepinephrine to mental stress observed in BN reflected a general reduction of sympathetic nervous system activity in this disorder. Lack of other sympathetic indices, however, prevented them from reaching any definitive conclusions. Therefore a second goal for this study was to assess an array of sympathetic measures in BN under both basal and stress conditions. The

J.H. Koo-Loeb et al. r Psychiatry Research 80 (1998) 13]27

final aim was to study non-hospitalized women with BN for sympathetic and neuroendocrine abnormalities, since previous studies have relied primarily on hospitalized women who may have refrained from their normal binge]purge behaviors for several weeks Že.g. Pirke et al., 1992b. 2. Method 2.1. Participants Participants were recruited from Chapel Hill and nearby communities through advertisements describing either a study investigating BN and blood pressure, or an investigation of stress and blood pressure Žfor control women.. Fifteen women met criteria for current bulimia nervosa, purging subtype, based on a structured clinical interview Žsee below. and were enrolled into the protocol. All women with BN had normal electrolytic panels Ždetermined at initial screening., were free of medication use and were free of any endocrine disorder or cardiovascular disease. Fifteen control women were matched for relevant demographic variables Žsee Table 1.. The control women also had normal levels of electrolytes, were free of medication usage and were free of any endocrine or cardiovascular disorder. Also, based on our desire to examine stress dysregulation in BN relative to healthy control women, control subjects had no current or prior psychiatric illness. 2.2. Initial screening and diagnostic inter¨ iew After obtaining informed consent, participants had blood drawn in order to insure that, for the BN group, electrolytic levels were within normal ranges. Next, participants completed a questionnaire pertaining to their health histories and then had 3]4 clinical blood pressures taken. A psychologist ŽS.G.. then conducted the Structured Clinical Interview for DSM-IV using criteria of the Diagnostic and Statistical Manual of Mental Disorders ŽAmerican Psychiatric Association, 1994, 4th ed.; SCID interview. for diagnoses related to current and prior history of Axis I psychiatric disorders. Owing to the primary focus of the study, which was to investigate sympatheticr

15

Table 1 Demographic and baseline variables Žmeans q S.D.. in women with bulimia nervosa and control subjects Variable

Sample group Bulimic Ž n s 15.

Demographics Age Žyears. Height Žinches. Weight Žpounds. Smokers Oral contraceptives Activity level kcalrweek

Control Ž n s 15.

25.3 Ž"7.03. 65.3 Ž"2.02. 130.9 Ž"15. 0 6

24.4 Ž"2.56. 64.9 Ž"2.52. 131.5 Ž"15. 1 5

2930 Ž"3.057.

1878 Ž"3.057.

adrenergic function using cardiovascular and other indices, only women with BN who engaged in self-induced vomiting as the compensatory behavior, and not any other form of purging Že.g. diuretics, laxatives., were included in this study. All determinations of Axis I diagnoses for women with BN and lack of psychiatric diagnoses for control subjects were based on a consensus diagnostic conference with a psychiatrist ŽC.P... 2.3. Experimental procedures 2.3.1. Physiological recording procedures in the laboratory Blood pressure was recorded non-invasively from the participant’s left arm, using a standard blood pressure cuff and electronic microphone. Three to four manual stethoscopic readings were taken using a sphygmomanometer to insure correct placement of the microphone. A laboratorybuilt semi-automated blood pressure monitor was used to determined systolic ŽSBP. and diastolic ŽDBP. blood pressure during the baseline and laboratory stressors. This system measured the blood pressure non-invasively using the auscultatory technique. The cuff pressure and Korotkoff sounds were recorded in analog form using the computerized Videograph system ŽDATAQ, Akron, OH, USA.. Calibration levels for pressure were entered before each study, and when replayed by a trained assistant, permitted the computer to provide precise levels of cuff pressure

16

J.H. Koo-Loeb et al. r Psychiatry Research 80 (1998) 13]27

corresponding to the onset Žsystolic pressure. and disappearance Ždiastolic pressure. of Korotkoff sounds. Impedance cardiography was used to permit non-invasive monitoring of stroke volume ŽSV. and heart rate ŽHR.. The Minnesota Impedance Cardiograph ŽModel 304B, Surcom, Minneapolis, MN, USA. was used in conjunction with tetrapolar band electrode configuration to record impedance d Zrdt and Z0 signals. Impedance and electrocardiogram signals were processed on-line and subsequently manually edited for accuracy by means of the interactive Cardiac Output Program software ŽBIT, Chapel Hill, NC, USA., which was developed and validated in our laboratory Žsee Sherwood et al., 1991.. For each minute of interest, a 40-s continuous sample of waveforms Žobtained concurrently with a blood pressure reading. was processed to generate an ensemble-averaged cardiac cycle from which SV was determined by means of the Kubicek et al. Ž1966. equation and HR was determined by the mean interbeat interval; cardiac output ŽCO. and total peripheral resistance ŽTPR. for these same minutes were then calculated using standard formulae Žsee Sherwood et al., 1990..

and averaged to constitute baseline levels. At the end of the seated baseline, blood was sampled from an antecubital vein for baseline epinephrine ŽEpi. and NE, for basal levels of thyroid hormones in order to assess metabolic state and for cortisol. Following this prestress baseline, mental stressors were presented, counterbalancing order within each group. 2.3.4. Paced Auditory Serial Addition Task This taped mental arithmetic task is designed to measure rate of information processing ŽGronwall, 1977. and is composed of four series of numbers from 1 to 9. Participants were instructed to add each number to the immediately preceding number, stating the answer aloud. Each series had progressively shorter interdigit intervals. Task duration was 9.5 min, and cardiovascular measures were taken once in each series and averaged to constitute task level. Rate of information processing is calculated as rate of digit presentationrnumber correct.

2.3.2. Laboratory protocol Participants were instructed to refrain from all over-the-counter medications for 24 h prior to testing and from caffeine on the day of testing, and to consume a light breakfast prior to visiting the laboratory. All testing occurred on a normal working or school day between 08.00 and 11.00 h and during days 2]11 of the menstrual cycle. Participants were first tested for ischemic pain sensitivity using the submaximal effort tourniquet procedure wsee Girdler et al. Ž1998a. for that reportx followed by a 10-min recovery period. Participants were then exposed to the following conditions: Seated Prestress Baseline, Paced Auditory Serial Addition Task and Interpersonal Speech Task.

2.3.5. Interpersonal Speech Task Participants were presented with a hypothetical situation involving a plausible real-life interpersonal hassle. Participants were instructed that they would have 2 min for Speech Preparation and that they would then be asked to give a 3-min speech describing what their actions and emotional responses would be in this situation. The hassle concerned an inconsiderate house-guest who was taking advantage of the participant’s hospitality. Stories were tape-recorded and subsequently replayed by three of the laboratory staff to be judged for poise, articulation and style. Bonus money was contingent on the judges’ ratings. Cardiovascular measures were taken at minute 2 of the Speech Preparation phase and at minutes 1 and 3 during the Speech phase. At the end of minute 1 of the Speech phase, blood was again sampled for stress-related Epi and NE.

2.3.3. Seated baseline A 10-min seated rest period preceded the mental stressors. During this time, cardiovascular measures were taken at Minutes 1, 3, 6, 8 and 10

2.3.6. Reco¨ ery An 8-min recovery period was imposed between the two stressors.

J.H. Koo-Loeb et al. r Psychiatry Research 80 (1998) 13]27

2.3.7. Post-task questionnaire A post-task questionnaire was administered immediately after each stressor, but prior to beginning the recovery period. A visual analogue scale was used to assess to what degree each experienced certain moods and emotions, including anger, tension, confusion, fatigue, ability to concentrate and inadequacy. In addition, the questionnaire also assessed the degree to which each felt helpless, satisfied with performance, the task to be fair, and the amount of effort exerted during the task. 2.3.8. Neuroendocrine assays Plasma levels of Epi and NE were determined by high performance liquid chromatography ŽHPLC.. The reliability of HPLC techniques for catecholamine assays has been well documented. All HPLC procedures were conducted at the UNC Hospitals General Clinical Research Center, where a state-of-the-art HPLC system has been set up specifically for catecholamine assays. The coefficient of variation for this plasma catecholamine assay Žbased on pooled control plasma. is under 10% and the sensitivity limit is 5 pgrml. Serum levels of triiodothyronine ŽT3 . were determined by radioimmunoassay ŽRIA. using commercial kits from ICN Pharmaceuticals. The T3 antiserum Žmeasuring total T3 s bound plus free. is highly specific for T3 , showing only 1]18% cross-reactivity with other thyroid hormones, with a sensitivity limit of 6.7 ngrdl. Plasma levels of cortisol were determined by RIA using commercial kits from ICN Pharmaceuticals. The sensitivity of the assay is excellent at 0.07 m grdl. The specificity of this RIA for cortisol is high, showing 0.05]2.2% cross-reactivity with similarly structured compounds, with the exception of prednisolone, where 94% cross-reactivity is obtained. 2.4. Psychosocial questionnaires The following questionnaires were given to participants to complete at home during the 24-h period before the test session: Ža. Beck Depression Inventory ŽBeck et al., 1961., a 21-item scale

17

designed to assess depressive symptomatology; Žb. Spielberger Trait Anxiety Inventory ŽSpielberger et al., 1970., a 20-item scale which measures a relatively stable characteristic of individuals to respond anxiously when faced with a stressful situation; Žc. Rosenberg’s Self-Esteem Scale ŽRosenberg, 1965., a 10-item scale which assesses self-esteem in terms of self-acceptance; Žd. Pearlin Mastery Scale ŽPearlin et al., 1981., which measures perceived control over their world; Že. Interpersonal Support Evaluation List ŽCohen et al., 1985., which assesses general perception of social support; Žf. Profile of Mood States ŽLorr and McNair, 1988., a 72-item scale with six bipolar subjective mood states including composed]anxious, agreeable]hostile, elated]depressed, confident]unsure, energetic]tired and clearheaded] confused; Žg. Perceived Stress Scale ŽCohen et al., 1993., a 14-item scale designed to measure the degree of stress perceived in life; Žh. Sarason Brief Social Support Questionnaire ŽSarason et al., 1987., which measures the number of available others the person feels that he or she can turn to in times of need and the person’s degree of satisfaction with the support given; and Ži. Ways of Coping Scale ŽLazarus and Folkman, 1984., a 60-item questionnaire assessing the variety of ways a person copes with stress in different situations, including problem-focusing, blaming others, counting blessings, problem-avoiding, blaming self, seeking social support, minimizing threat and engaging in wishful thinking. 2.5. Design and data analysis The first analytic approach involved examining group differences in baseline cardiovascular measures using a one-way analysis of variance ŽANOVA.. Next, group differences in cardiovascular and neuroendocrine responses to stress were analyzed using a 2 ŽGroup. = 3 ŽTask. repeated measures design. Owing to evidence that anxiety is associated with altered adrenergic function and cardiovascular responses Ž Hoehn-Saric and McLeod, 1988; Cameron et al., 1990. and because anxiety was the most prevalent current comorbid diagnosis among our women with BN Žsee Section

18

J.H. Koo-Loeb et al. r Psychiatry Research 80 (1998) 13]27

3., these analyses were run with and without covarying for Spielberger trait anxiety scores. A comparison of these effects is presented in Section 3. Where significant interactions emerged, subsequent simple effects analyses were conducted. Neuroendocrine data were analyzed using a 2 ŽGroup. = 2 ŽBaseline, Stress. ANOVA. Between-group t-tests were used to examine group differences in psychosocial measures.

quency, duration and intensity of weekly physical activities ŽAinsworth et al., 1993.1. 3.3. Group differences in stress reacti¨ ity Despite equivalent baseline cardiovascular measures, bulimic women showed evidence for blunted sympathetic reactivity to the mental stressors, regardless of whether analyses controlled for current anxiety levels or not.

3. Results 3.1. SCID } Axis I comorbidity The SCID revealed that the most prevalent comorbid disorders were Major Depressive Disorder ŽMDD. and Anxiety Disorders ŽAD., with Anxiety Disorders being the most prevalent current comorbidity. Specifically, 53% of our sample were diagnosed with a past history of MDD, while an additional 13% were diagnosed with current MDD. Past history of AD was diagnosed in 20% of our sample, while an additional 47% were diagnosed with current AD Žmostly obsessive]compulsive disorder.. A total of 26% of the bulimic women were diagnosed with either a past history or current Panic Disorder, 7% had current Substance Abuse Disorder Žalcohol., and 13% were diagnosed with a past history of anorexia, though no one had a current diagnosis of anorexia. 3.2. Demographic and baseline ¨ ariables Women with BN and control subjects did not differ on the relevant variables of age, height, weight, smoking status and oral contraceptive use Žsee Table 1.. Furthermore, no differences were evident between BN and control women in terms of baseline SBP Ž107 vs. 110 mmHg., DBP Ž66 vs. 69 mmHg., TPR Ž1049 vs. 1169 arbitrary units., CO Ž6.2 vs. 5.9 Lrmin., SV Ž100 vs. 92 ml., HR Ž63 vs. 65 bpm., or pre-ejection period ŽPEP. Ž118 vs. 115 ms., all P) 0.15. Lastly, consistent with the purging subtype of the bulimic women included in this protocol, groups did not significantly differ in weekly energy expenditure, which was calculated based upon self-reported fre-

3.3.1. Blood pressure and heart rate When analyses controlled for anxiety levels, women with BN showed marginally blunted SBP Žmain effect of Group: F1,26 s 3.92, Ps 0.058. and had significantly blunted HR reactivity Žmain effect of Group: F1,26 s 5.66, Ps 0.02. to all stressors compared with control women Žsee Figs. 1 and 2.. Though bulimic women also tended to exhibit lower DBP reactivity to stress, this difference was not statistically significant. When these analyses were repeated without controlling for anxiety, the magnitude of the blood pressure effects were enhanced ŽSBP and DBP main effects of Group: F1,27 s 4.63]6.89, P- 0.05., while the magnitude of difference in HR was diminished Žmain effect of Group: F1,27 s 2.91, P- 0.10.. 3.3.2. Pre-ejection period Consistent with the BP and HR effects, which suggested lesser sympathetic reactivity in BN, when analyses covaried for current anxiety, women with BN had greater PEP time intervals across all conditions Žmain effect of Group: F1,25 s 10.78, P- 0.01; see Fig. 3., indicating lesser sympathetic tone and reduced myocardial contractility during stressors ŽNewlin and Levenson, 1979; Caccioppo et al., 1994.. When analyses were

1

Although groups were not statistically different in kilocalories Žkcal. expended, analyses were conducted to confirm that differences in exercise could not account for group differences in cardiovascular responses to stress. For each dependent measure during stressors, a simultaneous multiple regression was performed which included first kilocalories and then group as independent variables. Kilocalorie expenditure was not a statistically significant independent predictor of stress responses for any measure.

J.H. Koo-Loeb et al. r Psychiatry Research 80 (1998) 13]27

19

Fig. 1. Systolic blood pressure Žadjusted means q S.D.. during seated prestress, preparation of speech, delivery of speech and math task in women with bulimia nervosa and control women.

repeated without controlling for anxiety, the magnitude of the group difference was reduced Žmain effect of Group: F1,27 s 2.83, Ps 0.10.. Taken together, these results suggest reduced sympathetic reactivity to stress in BN as reflected in BP, HR and PEP. Although controlling for anxiety levels influenced the magnitude of these effects, reducing BP differences but enhancing myocardial differences, the general pattern for blunted sympathetic reactivity in BN was evident with or without adjusting for current anxiety level. 3.3.3. Stroke ¨ olume, cardiac output and total peripheral resistance. There were no significant differences in SV, CO, or TPR between women with BN or control

women in response to any of the tasks, with or without controlling for anxiety levels. 3.3.4. Plasma norepinephrine and epinephrine Although there were no differences in absolute baseline Ž229.5 vs. 231.0 pgrml. or stress NE levels Ž224.6 vs. 262.7 pgrml. between BN women and control subjects, respectively, only the control women showed evidence for a significant increase in NE to the speech stress Ž Ps 0.07. relative to their own baseline level, while the women with BN showed no increase in NE to stress. More robust group differences were observed for Epi ŽFig. 4. where BN women showed lower plasma Epi levels during both baseline and speech stress compared with control women Žmain effect of

20

J.H. Koo-Loeb et al. r Psychiatry Research 80 (1998) 13]27

Fig. 2. Heart rate Žadjusted means q S.D.. during seated prestress, preparation of speech, delivery of speech, and math task in women with bulimia nervosa and control women.

Group: F1,25 s 9.19, Ps 0.01.. Similar to the patterns observed for NE, only the control women mounted a significant increase in Epi in response to stress Ž P- 0.02.. 3.3.5. Plasma cortisol Plasma samples obtained during the prestress baseline, which were meant to capture the cortisol response to the ischemic pain procedure that preceded mental stress ŽGirdler et al., 1998a., revealed that women with BN exhibited significantly greater plasma cortisol levels compared with control subjects Ž15.9" 1.6 m grdl vs. 9.5" 1.8 m grdl; F3,22 s 4.94, P- 0.05.. 3.3.6. Plasma triiodothyronine There was no difference in T3 levels between

bulimic women and control subjects Ž148 vs. 149 ngrdl, respectively.. 3.3.7. Neuropsychological performance measures for the Paced Auditory Serial Addition Task (PASAT) There were no group differences in PASAT performance, during any series. 3.3.8. Post-task questionnaires Group differences in moods and emotions during the tasks were evident. Analyses revealed that during the Speech task, bulimic women were more confused ŽT14,14 s 2.95, Ps 0.006., more frustrated ŽT14,14 s 2.05, P- 0.05., felt more inadequate ŽT14,14 s 3.18, Ps 0.004. and were more dissatisfied by their performance on the task ŽT14,14 s 2.53, Ps 0.02.. Despite the fact that

J.H. Koo-Loeb et al. r Psychiatry Research 80 (1998) 13]27

21

Fig. 3. Pre-ejection period Žadjusted means q S.D.. during seated prestress, preparation of speech, delivery of speech and math task in women with bulimia nervosa and control women.

women with BN and control women did not differ in performance during the math task ŽPASAT., bulimic women also felt more inadequate as compared with control subjects ŽT14,14 s 2.14, Ps 0.04. during that task. 3.4. Psychosocial questionnaires Psychosocial assessment revealed robust differences between women with BN and control subjects. Women with BN perceived more stress as compared with control subjects in the month prior to their laboratory testing ŽT14,14 s 3.75, P s 0.0008.. Bulimic women endorsed worse coping skills as compared with control women by avoiding their problems more ŽT13,14 s 4.40, P s

0.0002., blaming others more ŽT13,14 s 2.2, Ps 0.03., counting blessings less ŽT13,13 s y2.75, Ps 0.01., focusing less on the problem ŽT13,13 s y2.93, P s 0.009., blaming themselves more ŽT14,14 s 3.88, Ps 0.0006., seeking less social support ŽT14,13 s y3.48, Ps 0.0017., and engaging in significantly more wishful thinking ŽT14,14 s 3.9, Ps 0.0006., as compared with control women. Women with BN also had a lower sense of mastery over their world ŽT14,14 s y4.25, Ps 0.0002., had lower self-esteem ŽT14,14 s y4.46, P s 0.0001., and reported less social support in general, as compared with control subjects ŽT14,14 s y2.18, Ps 0.04., particularly less appraisal support ŽT14,14 s y2.23, Ps 0.03.. Moreover, bulimic women perceived significantly less support from

22

J.H. Koo-Loeb et al. r Psychiatry Research 80 (1998) 13]27

Fig. 4. Epinephrine Žadjusted means q S.D.. during seated prestress and during delivery of speech in women with bulimia nervosa and control women.

family ŽT14,12 s y2.29, Ps 0.03. and other people ŽT14,13 s y2.51, Ps 0.02., as compared with control subjects, and were less satisfied with what support they did have ŽT13,14 s y2.12, Ps 0.04.. The Profile of Mood States revealed that women with BN were significantly more hostile ŽT14,14 s y2.33, P s 0.03., more anxious ŽT14,14 s y3.52, Ps 0.002., more confused ŽT14,14 s y3.87, Ps 0.0006., more unsure of themselves ŽT14,14 s y4.29, P s 0.0002., more depressed ŽT14,14 s y2.69, Ps 0.01., and more tired ŽT14,14 s y2.02, Ps 0.05. than control subjects. Similarly, women with BN scored higher on the Spielberger Trait Anxiety Inventory ŽT14,14 s 4.36, Ps 0.0002. and on the Beck Depression Inventory Ž15.8 vs. 3.1: T14,14 s 4.17, Ps 0.0003..

4. Discussion Our results suggest that women with BN exhibit blunted sympathetic activation in response to mental stressors as compared with healthy control women. Specifically, though there were no differences in resting cardiovascular measures, after we controlled for the greater anxiety levels in BN, women with BN showed blunted systolic blood pressure and heart rate in response to all mental stressors compared with control women. Additionally, the pre-ejection period, a reliable indicator of sympathetic efferent activity on the heart and related to centrally mediated sympathetic tone ŽBerntson et al., 1994; Caccioppo et al., 1994., was elongated in women with BN,

J.H. Koo-Loeb et al. r Psychiatry Research 80 (1998) 13]27

reflecting lesser sympathetic innervation. Additionally, we found that BN was associated with significantly lower Epi levels both at rest and during mental stress, and although we observed no differences in basal NE between groups, control women mounted a marginally significant increase in NE during stress while women with BN showed no catecholamine reactivity. Differences in catecholamine levels in BN have been previously documented wGeorge et al. Ž1990., Pirke et al. Ž1992b.; see Pirke Ž1996. for reviewx. For example, while we are not aware of any studies examining Epi responses to mental stress in BN, reduced baseline Epi in BN was reported by George et al. Ž1990.. In addition, Pirke et al. Ž1992b. found that women with BN exhibited lower baseline NE and no significant NE response to a mental stressor as compared with a control group. Our results combined with others ŽPirke et al., 1992b; Pirke, 1996. suggest reduced sympathetic activation during stress in BN. However, we have extended the existing literature by examining a comprehensive set of sympathetic nervous system measures, both at rest and in response to stress. Taken together, these results suggest stress-induced sympathetic dysregulation in BN. It has been suggested that the observed adrenergic dysregulation in BN wsee Pirke Ž1996. for reviewx reflects, in part, metabolic deficiencies associated with reduced caloric intake ŽPirke et al., 1985, 1990, 1992a.. Indeed, many of the studies finding adrenergic differences included women with anorexia nervosa ŽPirke et al., 1992a. or BN women with reduced T3 levels Že.g. Pirke et al., 1985., which are indicative of significant caloric restriction. However, our sample of women with BN neither differed from control women in T3 levels nor in electrolytic profiles, as determined at initial screening. The women in our study were all of normal weight and did not engage in severe caloric restriction, as is consistent with the purging subtype diagnosis for BN employed in the present study. Furthermore, although starvation can have long-term consequences on adrenergic physiology, only 2r15 Ž13%. of the women with BN had a past history of anorexia nervosa. Thus, while speculative, our results may suggest that in

23

addition to the effects that severe diet or starvation can have on adrenergic function in eating disorders, adrenergic dysregulation may also be a predisposing factor in eating disorders, at least in BN, purging subtype, and not merely a consequence of altered eating patterns. While our interpretation is limited by the fact that we did not fully evaluate nutritional status in this study, we have no reason to believe that the women with BN were nutritionally deficient based upon the available evidence. It is not clear why BN might be associated with adrenergic dysregulation independent of nutritional status, but several possibilities exist. First, adrenergic differences may reflect the prevalence of mood disorders seen in BN Že.g. Pirke, 1996., and indeed, our comorbidity rates were similar to those in other reports ŽBrewerton et al., 1995; Garfinkel et al., 1995.. However, even when we controlled for the most prevalent current psychiatric condition, anxiety, analyses still revealed stress-induced sympathetic dysregulation in BN. Furthermore, only two of the women with BN were currently depressed at the time of testing. Thus, these results suggest that group differences in stress reactivity and sympathetic tone cannot be solely accounted for by the mood disorders associated with BN. An alternative explanation relates to the greater cortisol levels that we observed in the BN women following the ischemic pain testing that preceded mental stress testing. There is evidence that blunted sympathetic activity may be due to the suppressive effects of glucocorticoids ŽMunck and Naray-Fejes-Toth, 1992, 1994.. For example, bilateral adrenalectomy augments sympathoneural responses to immobilization stress in monkeys while corticosterone supplementation in adrenalectomized animals reduces the sympathetic response wsee Kvetnansky et al. Ž1995. for reviewx. However, the relationship between glucocorticoids and sympathetic]adrenal]medullary ŽSAM. activity may differ during psychological stress as opposed to physical stress ŽHellhammer and Wade, 1993; Komesaroff and Funder, 1994., and may differ depending on the species. For example, in human males, Malarkey et al. Ž1995. found that glucocorticoids did not suppress tonic or

24

J.H. Koo-Loeb et al. r Psychiatry Research 80 (1998) 13]27

phasic sympathetic activity, since dexamethasone decreased both basal and mental stress-induced ACTH and cortisol levels but did not affect catecholamine levels either at rest or in response to stress. Thus, the suppressive effects of glucosteroids on the sympathetic system may be less robust in humans during mental stress. Since we obtained cortisol only after physical stress, however, and not during mental stressors in the present study, it is unclear whether the blunted sympathetic activity seen in our sample of BN women was related to increased levels of cortisol or due to some other physiological andror psychological factorŽs.. Future studies in BN should assess cortisol and catecholamines simultaneously during mental stress in order to examine this issue. Another possibility for the dissociation that we observed between the SAM and hypothalamic]pituitary]adrenal ŽHPA. axes in BN may involve psychological and situational factors. Frankenhaeuser first proposed a model of Effort ŽStress. vs. Distress Ž1993. which has been supported in a series of subsequent studies wsee Frankenhaeuser Ž1993. for review; Frankenhaeuser et al., 1989; Lundberg et al., 1990x. Taken together, Frankenhaeuser’s studies have documented that when control and mastery over stressors are fostered, participants report positive feelings with high effort, a sense of mastery, active coping, and exhibit relative activation of the SAM axis with greater epinephrine secretion compared with cortisol secretion, and substantial increases in blood pressure and heart rate. When exposed to stressors where no control is possible, however, participants report distress in that they lack control, withdraw, report negative feelings with effort, and exhibit relative activation of the HPA axis with greater cortisol secretion as compared with epinephrine, and show reduced blood pressure and heart rate reactivity. The blunted blood pressure, heart rate, and epinephrine plus elongated PEP intervals Žindicating reduced myocardial sympathetic innervation. during mental stressors seen in the women with BN in the present study, coupled with their heightened cortisol response to pain stress, are consistent with this cognitive model. For example, despite similar performances of women with BN

and control subjects on the mental stressors, BN women reported feeling more confused, frustrated, inadequate, and dissatisfied with their performance during stressors. In addition, the psychosocial profile that we obtained for the women with BN is consistent with a state of distress since they perceived more stress, had worse coping skills, had lower sense of mastery and self-esteem, reported less social support, were less satisfied with the support they had, reported worse moods in general, experienced more anxiety, and were more depressed compared with our non-eating disordered control subjects. Thus, the cardiovascular, neuroendocrine and psychosocial profiles of the women with BN are consistent with a dissociation of the SAM and HPA axes, characteristic of ‘distress’ vs. ‘effort’. This is not to say, however, that SAM and HPA axis dissociation or blunted sympathetic responses to stress are unique to BN, since other clinical or chronically stressed groups have blunted stress responses or neuroendocrine axis abnormalities. However, the pattern of such dysregulation may confer pathophysiological specificity. For example, although cigarette smokers show blunted cardiac responses to stressors ŽGirdler et al., 1997., they also show blunted, and not heightened, cortisol reactivity ŽKirschbaum et al., 1993.. Regarding clinical populations, while women with severe premenstrual syndrome also show blunted heart rate and blood pressure during stress ŽGirdler et al., 1993., they exhibit elevated norepinephrine and blunted cortisol responses to stressors ŽGirdler et al., 1998b., the pattern opposite to what we observed in BN. Individuals with post-traumatic stress disorder ŽPTSD. also show SAM and HPA axis dissociation, but again, it is opposite to what we observed in BN since PTSD is associated with elevated NE to cortisol ratios ŽMason et al., 1988.. Finally, major depression is associated with increased, and not decreased, sympathetic output ŽStratakis and Chrousos, 1995., coupled with increased HPA activity ŽStratakis and Chrousos, 1995.. Clearly, more research is needed to investigate the pathophysiological significance of different patterns of adrenergic and HPA axis dysregulation during stress, but our results suggest that the pattern of dysregulation that we observed in women with BN may be unique to this disorder

J.H. Koo-Loeb et al. r Psychiatry Research 80 (1998) 13]27

and may relate to the negative psychosocial profile associated with ‘distress’ in combination with a certain biological vulnerability. In conclusion, we found that women with BN demonstrated blunted sympathetic activation during mental stressors, as reflected in reduced heart rate, blood pressure, and epinephrine coupled with elongated PEP intervals relative to findings in control women. In contrast, women with BN showed elevated plasma cortisol levels compared with control subjects. Despite equivalent performance during stressors, women with BN also responded with significantly more negative affect during the tasks, and exhibited more negative affect, greater perceived stress and poorer coping skills in their daily lives relative to the control women. One obvious limitation to our study is its relatively small sample size. Thus, these results should be considered preliminary. Nevertheless, we hypothesize that the combination of the observed physiological and psychological response patterns reflects a dissociation of the SAM and HPA axes, characteristic of a state of distress as opposed to active effort coping when confronted with interpersonal challenge. Of course, our study cannot speak to the important issue of whether the dissociation of the SAM and HPA axes during stress precedes or is a consequence of BN, nor can it speak to the temporal relationship between the negative coping styles and affect and the dampened sympathetic activation since all measures in the present study were obtained after BN had been well established. Future studies assessing these measures in individuals at risk for developing BN, such as those scoring high on an eating disorder scale but who have never met the frequency or severity criteria for BN, may help to elucidate this issue. Regardless of the causal relationship, however, the results of our study add to existing evidence that individuals with BN would benefit from interventions designed to improve coping skills and self-perceptions, and it would be of great scientific and clinical interest to examine whether such cognitive shifts normalize the stress response pattern seen in BN. Acknowledgements This research was supported by National Insti-

25

tutes of Health Grants, R01-MH51246, GCRCRR00046, and MHCRC-MH33127, and by the Foundation of Hope for Research and Treatment of Mental Illness. The authors would like to thank Catherine Stanwyck for her invaluable assistance in data management and analysis. In addition, the authors would like to thank all participants for the unselfish time and effort that they devoted to making this study possible. References Abraham, H., Joseph, A., 1987. Bulimic vomiting alters pain tolerance and mood. International Journal of Psychiatry in Medicine 16, 311]316. Ainsworth, B., Haskell, W., Leon, A., Jacobs, D., Montoye, H., Sallis, J.F., Paffenbarger, R., Jr., 1993. Compendium of Physical Activities: classification of energy costs of human physical activities. Medicine and Science in Sports and Exercise, 25, 71]80. American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental Disorders, 4th ed. APA, Washington, D.C. Barrios, B., Pennebaker, J., 1982. Bulimia Nervosa: Descriptive Data and a Conceptual Framework. Presented at the Annual Meeting of the Association for the Advancement of Behavioral Therapy, Los Angeles, CA. Beck, A., Ward, C., Mendelson, M., Mock, J., Erbaugh, J., 1961. An inventory for measuring depression. Archives of General Psychiatry 4, 561]571. Berntson, G., Cacioppo, J., Binkley, P., Uchino, B., Quigley, K., Fieldstone, A., 1994. Autonomic cardiac control. III. Psychological stress and cardiac responses in autonomic space as revealed by pharmacological blockades. Psychophysiology 31, 599]608. Brewerton, T., Lydiard, R., Herzog, D., Brotman, A., O’Neil, P., Ballenger, J., 1995. Comorbidity of axis I psychiatric disorders in bulimia nervosa. Journal of Clinical Psychiatry 56, 77]80. Briere, J., 1989. Symptomatology associated with childhood sexual victimization in a nonclinical adult sample. Child Abuse and Neglect 12, 51]59. Caccioppo, J., Berntson, G., Binkley, P., Quigley, K., Uchino, B., Fieldstone, A., 1994. Autonomic cardiac control. II. Noninvasive indices and basal response as revealed by autonomic blockades. Psychophysiology 31, 586]598. Cameron, O., Smith, C., Lee, M., Hollingsworth, P., Hill, E., Curtis, G., 1990. Adrenergic status in anxiety disorders: platelet alpha 2-adrenergic receptor binding, blood pressure, pulse, and plasma catecholamines in panic and generalized anxiety disorder patients and in normal subjects. Biological Psychiatry 28, 3]20. Cattanach, L., Rodin, J., 1988. Psychosocial components of stress process in bulimia. International Journal of Eating Disorders 7, 75]88.

26

J.H. Koo-Loeb et al. r Psychiatry Research 80 (1998) 13]27

Cattanach, L., Malley, R., Rodin, J., 1988. Psychologic and physiologic reactivity to stressors in eating disordered individuals. Psychosomatic Medicine 50, 591]599. Chrousos, G.P., 1992. Regulation and dysregulation of the hypothalamic]pituitary]adrenal axis: the corticotropin-releasing hormone perspective. Endocrinology and Metabolism Clinics of North America 21, 933]858. Chrousos, G.P., 1995. The hypothalamic]pituitary]adrenal axis and immune-mediated inflammation. New England Journal of Medicine 332, 1351]1362. Chrousos, G.P., Gold, P.W., 1992. The concepts of stress and stress system disorders. JAMA 267, 1244]1252. Cohen, S., Kamarck, T., Mermelstein, R., 1983. A global measure of perceived stress. Journal of Health and Social Behavior 24, 385]396. Cohen, S., Mermelstein, R., Kamarck, T., Hoberman, H., 1985. Measuring the functional components of social support. In: Sarason, I.G., Sarason, B.R. ŽEds.., Social Support: Theory, Research and Applications. Martinus Nijhoff, Dordrecht, pp. 73]94. Copeland, P., Herzog, D., Carr, D., Klibanski, A., MacLaughlin, R., Martin, J., 1988. Effect of dexamethasone on cortisol and prolactin responses to meals in bulimic and normal women. Psychoneuroendocrinology 13, 273]278. Crowther, J., Chernyk, B., 1986. Bulimia and binge eating in adolescent females: a comparison. Addictive Behavior 11, 415]424. Frankenhaeuser, M., 1983. The sympathetic]adrenal and pituitary]adrenal response to challenge: comparison between the sexes. In: Dembroski, T.M., Schmidt, T., Blumchen, G. ŽEds.., Biobehavioral Bases of Coronary Heart Disease. Karger, New York, pp. 91]105. Frankenhaeuser, M., Hedman, M., Bergman-Losman, B., Wallin, L., 1989. Stress on and off the job as related to sex and occupational status in white-collar workers. Journal of Organizational Behavior 10, 321]346. Garfinkel, P., Lin, E., Goering, P., Spegg, C., Goldbloom, D., Kennedy, S., Kaplan, A., Woodside, D., 1995. Bulimia nervosa in a Canadian community sample: prevalence and comparison of subgroups. American Journal of Psychiatry 152, 1052]1058. George, D., Kaye, W., Goldstein, D., Brewerton, T., Jimerson, D., 1990. Altered norepinephrine regulation in bulimia: effects of pharmacological challenge with isoproterenol. Psychiatry Research 33, 1]10. Girdler, S., Pedersen, C., Stern, R., Light, K., 1993. Menstrual cycle and premenstrual syndrome: modifiers of cardiovascular reactivity in women. Health Psychology 12, 180]192. Girdler, S., Jamner, L., Jarvik, M., Soles, J., Shapiro, D., 1997. Smoking status and nicotine administration differentially modify hemodynamic stress reactivity in men and women. Psychosomatic Medicine 59, 294]306. Girdler, S., Koo-Loeb, J., Pedersen, C., Brown, H., Maixner, B., 1998a. Blood pressure-related hypoalgesia in bulimia nervosa. Psychosomatic Medicine Žin press.. Girdler, S., Pedersen, C., Straneva, P., Leserman, J., Stanwyck, C., Benjamin, S., Light, K., 1998b. Dysregulation of cardiovascular and neuroendocrine responses to stress in pre-

menstrual dysphoric disorder. Psychiatry Research Žin press.. Gronwall, D., 1977. Paced auditory serial-addition task: a measure of recovery from concussion. Perceptual and Motor Skills 44, 367]373. Hellhammer, D., Wade, S., 1993. Endocrine correlates of stress vulnerability. Psychotherapy and Psychosomatics 60, 8]17. Hoehn-Saric, R., McLeod, D., 1988. The peripheral sympathetic nervous system. Psychiatric Clinics of North America 11, 375]386. Hudson, J., Laffer, P., Pope, H., 1982. Bulimia related to affective disorder by family history and response to the dexamethasone suppression test. American Journal of Psychiatry 139, 685]687. Kirschbaum, C., Strasburger, J., Langkrar, J., 1993. Attenuated cortisol response to psychological stress but not to CRH or ergometry in young habitual smokers. Pharmacology, Biochemistry and Behavior 44, 527]531. Kling, M.A., Perini, G.I., Demitrack, M.A., Geracioti, T.D., Linnoila, M., Chrousos, G.P., Gold, P.W., 1989. Stress-responsive neurohormonal systems and the symptom complex of affective illness. Psychopharmacology Bulletin 25, 312]318. Komesaroff, P.A., Funder, J.W., 1994. Differential glucocorticoid effects on catecholamine responses to stress. American Journal of Physiology 266, E118]E128. Kubicek, W., Karnegis, J., Patterson, R., Witsoe, D., Mattson, R., 1966. Development and evaluation of an impedance cardiac output system. Aerospace Medicine 37, 1208]1212. Kvetnansky, R., Pacak, K., Fukuhara, K., Viskupic, E., Hiremagalur, B., Nankova, B., Goldstein, D., Sabban, E., Kopin, I., 1995. Sympathoadrenal system in stress. Interaction with the hypothalamic]pituitary]adrenocortical system. Annals of the New York Academy of Sciences 771, 131]158. Lazarus, R., Folkman, S., 1984. Stress, Appraisal and Coping. Springer Publishing Company, New York, pp. 328]333. Lingswiler, V., Crowther, J., Stephens, M., 1989. Affective and cognitive antecedents to eating episodes in bulimia and binge eating. International Journal of Eating Disorders 8, 533]539. Lorr, M., McNair, D., 1988. Profile of Mood States: Bipolar Form. Educational and Industrial Testing Service, San Diego, CA. Lundberg, U., Melin, B., Fredrikson, M., Tuomisto, M., Frankenhaeuser, M., 1990. Comparison of neuroendocrine measurements under laboratory and naturalistic conditions. Pharmacology, Biochemistry and Behavior 37, 697]702. Malarkey, W., Lipkus, I., Cacioppo, J., 1995. The dissociation of catecholamine and hypothalamic]pituitary]adrenal responses to daily stressors using dexamethasone. Journal of Clinical Endocrinology and Metabolism 80, 2458]2463. Mason, J.W., Giller, E.L., Kosten, T.R., Harness, L., 1988. Elevation of urinary norepinephrinercortisol ratio in posttraumatic stress disorder. Journal of Nervous and Mental Disease 176, 498]502. Mortola, J., Rasmussen, D., Yen, S., 1989. Alterations of the adrenocorticotropin]cortisol axis in normal weight bulimic

J.H. Koo-Loeb et al. r Psychiatry Research 80 (1998) 13]27 women: evidence for a central mechanism. Journal of Clinical Endocrinology and Metabolism 68, 517]522. Munck, A., Naray-Fejes-Toth, A., 1992. The ups and downs of glucocorticoid physiology: permissive and suppressive effects revisited. Molecular and Cellular Endocrinology 90, 115]130. Munck, A., Naray-Fejes-Toth, A., 1994. Glucocorticoids and stress: permissive and suppressive actions. Annals of the New York Academy of Sciences 746, 115]130; Discussion 131]133. Newlin, D., Levenson, R., 1979. Pre-ejection period: measuring beta-adrenergic influences upon the heart. Psychophysiology 16, 546]553. Pearlin, L., Menaghan, E., Lieberman, M., Mullan, J., 1981. The stress process. Journal of Health and Social Behavior 22, 337]356. Pirke, K., 1996. Central and peripheral noradrenalin regulation in eating disorders. Psychiatry Research 62, 43]49. Pirke, K., Pahl, J., Schweiger, U., Warnhoff, M., 1985. Metabolic and endocrine indices of starvation in bulimia: a comparison with anorexia nervosa. Psychiatry Research 15, 33]39. Pirke, K., Tuschl, R., Spyra, B., Laessle, G., Schweiger, U., Broocks, A., Sambauer, S., Zitzelsberger, G., 1990. Endocrine findings in restrained eaters. Physiology and Behavior 47, 903]906. Pirke, K., Kellner, M., Philipp, E., Laessle, R., Krieg, J.,

27

Fichter, M., 1992a. Plasma norepinephrine after a standardized test meal in acute and remitted patients with anorexia nervosa and in healthy controls. Biological Psychiatry 31, 1074]1077. Pirke, K., Platte, P., Laessle, R., Seidl, M., Fichter, M., 1992b. The effect of a mental challenge test of plasma norepinephrine and cortisol in bulimia nervosa and in controls. Biological Psychiatry 32, 202]206. Rosenberg, M., 1965. Society and the Adolescent Self-image. Princeton University Press, Princeton, NJ. Sarason, I., Sarason, B., Shearin, E., Pierce, G., 1987. A brief measure of social support: practical and theoretical implications. Journal of Social and Personal Relationships 4, 497]510. Sherwood, A., Dolan, C., Light, K., 1990. Hemodynamics of blood pressure responses during active and passive coping. Psychophysiology 27, 656]668. Sherwood, A., Carter, L., Jr., Murphy, C., 1991. Cardiac output by impedance cardiography: two alternative methodologies compared with thermodilution. Aviation Space and Environmental Medicine 62, 116]122. Spielberger, C., Gorsuch, R., Lushene, R., 1970. STAI Manual for the State-Trait Anxiety Inventory. Consulting Psychologists Press, Palo Alto, CA. Stratakis, C., Chrousos, G., 1995. Neuroendocrinology and pathophysiology of the stress system. Annals of the New York Academy of Sciences, 771, 1]18.