Psychosomatic Medicine
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowRequest Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Koo-Loeb, J. H.
Right arrow Articles by Girdler, S. S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Koo-Loeb, J. H.
Right arrow Articles by Girdler, S. S.
Related Collections
Right arrow Neuroendocrine
Right arrow Eating Disorder
Psychosomatic Medicine 62:539-548 (2000)
© 2000 American Psychosomatic Society

ORIGINAL ARTICLES

Women With Eating Disorder Tendencies Display Altered Cardiovascular, Neuroendocrine, and Psychosocial Profiles

Jeannie H. Koo-Loeb, PhD, Nancy Costello, PhD, Kathleen C. Light, PhD and Susan S. Girdler, PhD

From the Departments of Psychology (J.H.K.-L., K.C.L., S.S.G.), Dental Ecology (N.C.), and Psychiatry (K.C.L., S.S.G.), University of North Carolina at Chapel Hill, Chapel Hill, NC.

Address reprint requests to: Susan S. Girdler, PhD, Department of Psychiatry, CB 7175, University of North Carolina, Chapel Hill, NC 27599-7175.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 NOTES
 ACKNOWLEDGMENTS
 REFERENCES
 
OBJECTIVE: Cardiovascular, neuroendocrine, and psychosocial profiles were investigated in women with eating disorder tendencies, but who had never met clinical criteria for an eating disorder, and in healthy controls.

METHODS: Twenty-six women who scored in the highest distribution of the Eating Disorder Inventory bulimia subscale (HEDI women) and 27 women who scored in the lowest distribution (LEDI women) completed psychosocial questionnaires, underwent a speech reactivity task for measures of blood pressure and heart rate reactivity, and also underwent 24-hour ambulatory blood pressure monitoring and urinary neuroendocrine collection.

RESULTS: The HEDI women exhibited increased blood pressure and heart rate reactivity to the speech task and increased 24-hour urinary cortisol, but decreased 24-hour urinary norepinephrine compared with LEDI women. There were no overall group differences in 24-hour ambulatory blood pressure levels, but negative mood and tension were associated with greater systolic blood pressures for all women. Finally, HEDI women reported greater depressive symptoms and anxiety, lower self-esteem and sense of mastery, less social support, poor coping skills, and greater emotional impact of daily stressors relative to LEDI women.

CONCLUSIONS: These results indicate that the same pattern of neuroendocrine and psychosocial profiles seen in prior studies of bulimia nervosa are also present in women with eating disorder tendencies.

Key Words: eating disorders • bloodpressure • norepinephrine • cortisol • psychological profiles • ambulatoryblood pressure monitoring

Abbreviations: HEDI = High Eating Disorder Inventory women; LEDI = LowEating Disorder Inventory women; BN = bulimia nervosa; NE =norepinephrine; EPI = epinephrine; SAM =sympathetic-adrenal-medullary; HPAC =hypothalamic-pituitary-adrenal-cortical; ABPM = ambulatory bloodpressure monitoring; EDI = Eating Disorder Inventory; SCID-NP= Structured Clinical Interview, nonpatient edition; DA =dopamine; RIA = radioimmunoassay; HPLC = high performanceliquid chromatography; SBP = systolic blood pressure; DBP =diastolic blood pressure; MAP = mean arterial pressure; HR =heart rate; ANOVA = analysis of variance; ANCOVA = analysisof covariance; VAS = Visual Analogue Scale.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
METHODS
 RESULTS
 DISCUSSION
 NOTES
 ACKNOWLEDGMENTS
 REFERENCES
 
The eating disorder bulimia nervosa is estimated to afflict from 1% to 4% of the general female population (1–6), but as many as 5% to 10% of college students (7) and up to 14% of whites (7). Recent evidence indicates that the prevalence for BN is on the increase (1). Like other psychological disorders, there is evidence that stress plays a major role. Stress, particularly interpersonal stress (8), has been implicated in the etiology (9, 10) as well as maintenance (6, 11) of this disorder. In addition, BN seems to be associated with a negative psychosocial profile, with studies reporting increased depression (10, 12–15), anxiety (11, 13, 14, 16–18), and decreased self-esteem (18, 19) in bulimic women.

Despite the evidence for increased life stress in BN, only a small handful of studies have investigated stress-induced physiological measures in BN (20–22). Pirke et al. (22) found that compared with controls, women with BN had significantly increased levels of cortisol but decreased plasma NE levels during control conditions, whereas they showed a blunted response for both cortisol and NE during mental stress. These results are similar to our prior observations (21) where, compared with controls, bulimic women showed significantly lower plasma EPI levels during baseline and in response to a speech stressor and a nonsignificant increase in plasma NE to the stressor. On the other hand, whereas Pirke et al. (22) found decreased cortisol response to mental stress, Girdler et al. (20) found increased cortisol response to an ischemic pain task in women with BN. This discrepancy in cortisol response to stress may be a result of differences in neuroendocrine responses to a mental stressor, which elicits active coping mechanisms (23), vs. a physical stressor, which may elicit passive coping mechanisms (24).

Only two studies of which we are aware have examined stress-induced cardiovascular responses in women with eating disorders. In the first, Cattanach et al. (11) found no differences in blood pressure and pulse rate responses to mental stress between women who scored high on an eating disorder questionnaire vs. those who scored low. In our recent investigation (21), however, we found blunted blood pressure and heart rate reactivity to speech and math stress in bulimic women relative to controls. The discrepancy between these two studies may be due to differences in the samples investigated. The Cattanach et al. (11) sample was based on questionnaire assessment with no verification of eating disorder status via structured interview and thus may have included a very heterogeneous group. Our study sample (21), on the other hand, was composed exclusively of women who met strict DSM-IV criteria for BN.

Considering the available evidence for cardiovascular and neuroendocrine differences between bulimic women and controls, the results suggest dysregulation in SAM and HPAC axes in women with established BN. However, what remains unclear is whether the neuroendocrine and cardiovascular dysregulation, as well as the negative psychosocial profiles consistently observed in women with established BN, represent a consequence of their eating disorder or whether they represent markers for increased vulnerability to developing an eating disorder. Although longitudinal studies would be needed to fully address this issue, neuroendocrine and psychosocial assessment in women with bulimic tendencies but who have never met criteria for BN may provide preliminary insight into the progression of disturbances associated with the development of BN or other eating disorders.

Thus, the purpose of this study was to compare women who scored in the upper distribution of an eating disorder scale, but who had never met criteria for an eating disorder, with controls scoring in the lowest distribution of the eating disorder scale. Groups were compared for cardiovascular reactivity to mental stress, 24-hour urinary neuroendocrine measures, and psychosocial profiles. In addition, to increase validity over previous laboratory-based stress studies, this study also investigated blood pressure responses in the natural environment using ABPM. Specifically, via the use of diary records of mood and interpersonal interactions, our goal was to assess the degree to which these variables differentially influence blood pressure in women with eating disorder tendencies. We hypothesized that if neuroendocrine, cardiovascular, and psychosocial differences exist before the development of BN, we would observe evidence for decreased sympathetic activation but increased HPAC (ie, cortisol) activation, as well as a negative psychosocial profile in women with eating disorder tendencies compared with controls. In addition, if there are differences in ambulatory blood pressures between groups, we predicted that negative interpersonal interactions and negative mood, in particular, would be associated with the greatest blood pressure differences between groups.


   METHODS
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
RESULTS
 DISCUSSION
 NOTES
 ACKNOWLEDGMENTS
 REFERENCES
 
Subjects
Fifty-three women between the ages of 18 and 30 served as participants. All participants were free of any established medical conditions, including cardiovascular and neuroendocrine disorders and were not taking any prescription medication except oral contraceptives, as indicated by the self-report health history questionnaire. Blood pressure assessments confirmed that all participants were normotensive. Participants were paid for their participation.

Recruitment and Initial Screening
All participants were recruited through initial screening stations set up on a large university campus. Initial screening efforts involved 185 women who, after obtaining informed consent, completed a brief health history questionnaire and the EDI (25). The EDI is a 64-item questionnaire that includes eight subscales designed to tap into both the behavioral and cognitive components of eating disorders. The EDI has a test-retest correlation of 0.96 (over a period of 3 weeks) (26) and a correlation of 0.66 with the Eating Attitudes Test, another commonly used screening measure for eating disorders (27). Untransformed data were used to score the EDI, as recommended for a nonclinical population (28). To enhance comparability with our previous study that included women meeting DSM-IV criteria for BN, the bulimia subscale was selected as the scale with which to identify the high and low eating disordered tendency groups. This subscale assesses a tendency toward episodes of uncontrollable over-eating that may or may not be followed by self-induced vomiting. Women whose scores fell into approximately the top and bottom quintiles of the EDI bulimia scale (and where there were clear breaks in score distribution) were defined as our high (N = 31) and low (N = 27) eating disordered tendency groups (HEDI and LEDI, respectively). Scores ranged from 9 to 26 in the HEDI group and from 0 to 3 in the LEDI group. The maximum possible score on this subscale is 35.

Diagnostic Interview
The 31 HEDI and the 27 LEDI women were then brought in for a SCID-NP, based on DSM-IV criteria, to assess current and past Axis I disorders. The SCID-NP interviews were conducted by a clinical psychologist (N.C.). The interviews were conducted to confirm that neither group had any current Axis I diagnosis (including an eating disorder) nor had met criteria for an eating disorder diagnosis in the past. Five women in the HEDI group met criteria for one or more exclusionary Axis I diagnoses. None of the LEDI women met criteria for any current Axis I disorder or past eating disorder. This resulted in a final sample of 26 women in the HEDI group and 27 in the LEDI group.

Psychosocial Measurement
After completing the SCID, participants were asked to complete the following questionnaires at home: Beck Depression Inventory (29), a 21-item scale designed to assess depressive symptomatology with 11 cognitive, two overt behavior, five somatic, and one interpersonal question; Spielberger Trait Anxiety Inventory (30), a 20-item scale that measures a relatively stable characteristic of individuals to respond anxiously when faced with a stressful situation; Rosenberg’s Self-Esteem Scale (31), a 10-item scale that assesses self-esteem in terms of self-acceptance; Pearlin Mastery Scale (32), which measures perceived control over one’s environment; Interpersonal Support Evaluation List (33), which assesses general perception of social support and includes four 10-item scales measuring "appraisal" (perceived availability of someone who could listen to one’s problems), "self-esteem" (perceived positive comparison of self with others), "belonging" (perceived availability of people to do things with), and "tangibility" (perceived availability of material aid); Profile of Mood States (34), a 72-item scale with six bipolar subjective mood states including composed-anxious, agreeable-hostile, elated-depressed, confident-unsure, energetic-tired, and clearheaded-confused; Perceived Stress Scale (35), a 14-item scale designed to measure the degree of stress that one perceives in his or her life in the past month; Sarason Brief Social Support Questionnaire (36), 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; Ways of Coping Scale (37), a 60-item questionnaire assessing the variety of ways a person copes with stress in varying situations, including problem-focusing, blaming others, counting blessings, problem-avoiding, blaming self, seeking social support, minimizing threat, and engaging in wishful thinking; and the Daily Stress Inventory (38), used to assess both the frequency and emotional impact of stressful events occurring during the 24-hour period of blood pressure assessment. We also assessed frequency, duration, and intensity of weekly physical activities, including exercise, and then calculated average weekly kilocalories expended (39).

Ambulatory Monitoring
Participants were instructed to avoid medication, caffeine, and alcohol 24 hours before ABPM and to eat a light breakfast on the day of monitoring. On the day of ABPM and 24-hour urine collection (see below), participants were fitted with the custom designed and built Accutracker-II monitor (SunTech, Raleigh, NC). The Accutracker has been well validated against intraarterial measures of blood pressure (40).

All participants were fitted with the Accutracker-II between 8 AM and 11 AM by an experienced researcher (J.K.). The monitor requires the placement of three EKG leads on the upper body, a piezoelectric microphone positioned over the brachial artery, and a standard blood pressure cuff on the nondominant arm. After instrumentation, a minimum of six stethoscopic blood pressure readings (three seated and three standing) were taken simultaneously with six automated readings. Light et al. (40) have shown that although automated blood pressure readings are highly correlated with stethoscopic readings, automated readings, including those of the blood pressure monitor that we used, cannot be assumed to reflect absolute stethoscopic levels. Thus, we used standard methods (41) to equate the data determined by the two methods by applying an individually derived correction factor to each automated reading.

All monitoring occurred on a typical class or workday as identified by the participant. The monitor was programmed to automatically initiate four blood pressure readings at random times each hour during waking hours and two readings per hour during sleep. If an error was detected (eg, arm movement), the monitor was programmed to take an additional reading exactly 2 minutes later. During waking hours only, every time the blood pressure cuff inflated, participants were instructed to complete a brief diary page describing posture, activity, location, consumption of medication or caffeine, others present, physical exertion, and level of stress. Satisfaction with interpersonal interactions, mental activity (ie, inactive vs. busy), emotional state (ie, negative vs. positive mood; relaxed vs. tense), urge to eat (ie, not at all vs. extreme), and hunger (ie, not at all vs. extreme) were assessed using visual analogue scales.

Interpersonal Speech Task
After ABPM instrumentation, but before leaving the laboratory, participants underwent a controlled speech task based on the speech stressor of Saab et al. (42). Participants were read a hypothetical situation involving an interpersonal hassle. The hassle concerned an inconsiderate relative who was staying in the home of the participant and taking advantage of the participant’s hospitality. Each was instructed that they would have 2 minutes for speech preparation and that they would then be asked to give a 3-minute speech describing what they would say to the relative, what actions they would take and how they thought the relative would respond. Also, each was asked to describe any emotions she might feel during this situation. Participants were instructed that the stories would be tape-recorded and subsequently replayed by three of the laboratory staff to be judged for poise, articulation, and style. Cardiovascular measures were taken at minute 2 of the speech preparation and at minutes 1 and 3 during the speech.

Urinary Neuroendocrine Assessment
After the speech stressor and before leaving the laboratory, participants were provided with urine collection containers and instructed on the timing and collection of the 24-hour urine sampling to coincide with the 24-hour period of ABPM. Urinary cortisol, NE, EPI, and DA were measured to examine circulating indices of HPAC vs. SAM axis activation. Urinary cortisol concentration was measured by RIA using commercial kits from ICN Biomedicals, Inc. All RIA procedures were conducted at University of North Carolina’s Mental Health Clinical Research Center. The sensitivity of the assay is 0.07 µg/dl.

Urinary NE, EPI, and DA concentrations were determined using HPLC. 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 urinary catecholamine assay is <10% and the sensitivity limit is ca. 5 ng/ml.

Statistical Analyses
Potential group differences in demographic variables, cardiovascular baseline levels, energy expenditure, psychosocial measures, diary ratings, and 24-hour urinary neuroendocrine measures were analyzed using ANOVA. Group differences in ethnicity and oral contraceptive use were analyzed using {chi}2 analyses. Analyses regarding cardiovascular reactivity to the speech were based on change scores, defined as mean task level minus mean baseline level. Baseline BP and HR was determined by averaging a minimum of three seated BP and HR readings taken at least 1 minute apart during the initial instrumentation period (see above). Anticipatory responses during speech preparation were not included in the analyses. Because groups did not differ in any seated baseline measure (Table 1), group differences in cardiovascular reactivity to speech were analyzed using a one-way ANOVA and did not covary for baseline levels.


View this table:
[in this window]
[in a new window]
  		Table 1. Demographic and Baseline Variables (Mean + SEM) in Women with High (HEDI) and Low (LEDI) Eating Disorder Inventory Scores
 
Group differences in absolute blood pressure levels during the course of the 24-hour day were analyzed in the following ways. First, differences in overall 24-hour blood pressures were analyzed using ANCOVA, with posture (ie, standing, seated, and supine) as the covariate. Second, mean blood pressures were calculated per group per time period (ie, daytime, evening, sleep, and morning) and analyzed using a 2 (group: HEDI vs. LEDI) x 4 (time: daytime, evening, sleep, and morning) repeated measure ANCOVA with posture as the covariate and time as the repeated factor. To examine group differences in blood pressure levels as a function of satisfaction with interpersonal interactions, mood states (ie, happy/sad and tense/relaxed), urge to eat, and hunger (measured by VAS in the diary) were divided to create high and low ends for each VAS (ie, VAS ratings were split down the middle). To increase reliability for any particular VAS, it was required that the participant endorse both high and low ratings at least 3 times in the 24-hours to be included in the analyses. The high and low ends of each VAS were then treated as within-subjects repeated measures. Blood pressure levels for each diary measure were then analyzed using a 2 (group: HEDI vs. LEDI) x 2 (rating: high vs. low VAS ratings) repeated measures analysis of covariance, with posture as the covariate.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
DISCUSSION
 NOTES
 ACKNOWLEDGMENTS
 REFERENCES
 
Eating Disorders Inventory
Although the groups were defined based on the bulimia subscale of the EDI, the HEDI and LEDI groups were significantly different on five of the eight subscales. Specifically, the HEDI group scored higher than the LEDI group on drive for thinness (18.8 vs. 8.1; F(1,51) = 39.01, p = .0001), interoceptive awareness (20.6 vs. 9.3; F(1,51) = 30.39, p = .0001), bulimia (13.1 vs. 1.7; F(1,51) = 191.73, p = .0001), body dissatisfaction (29.3 vs. 17.7; F(1,51) = 16.96, p = .0001), and ineffectiveness (17.5 vs. 9.03; F(1,51) = 30.98, p = .0001) subscales. There were no differences on the maturity fears (15.2 vs. 12.6), perfectionism (19.3 vs. 17.3), or interpersonal distrust (12.04 vs. 10.5) subscales. Thus, these results suggest that the HEDI group had significant tendencies toward cognitive and behavioral eating abnormalities.

Demographic and Baseline Cardiovascular Variables
Analyses revealed that groups did not differ in day of menstrual cycle tested, weight, height, caffeine intake, alcohol intake, smoking, education level, age, oral contraceptive use, or energy expenditure (Table 1). However, as expected, significantly fewer African Americans fell into the HEDI group ({chi}2(2, N = 53) = 13.7, p = .001)1. Groups did not differ in seated or standing stethoscopic SBP, DBP, MAP, or HR (see Table 1 for seated values).

Cardiovascular Responses to Acute Interpersonal Speech Stress
Regarding cardiovascular reactivity to speech stress, the HEDI group had higher SBP reactivity (F(1,49) = 6.8, p = .01), DBP reactivity (F(1,49) = 6.85, p = .01), MAP reactivity (F(1,49) = 9.07, p < .01), and HR reactivity (F(1,48) = 4.35, p = .04) than the LEDI group (Figure 1).



View larger version (14K):
[in this window]
[in a new window]
  		Fig. 1. Interpersonal speech task reactivity (means + SEM) in women with high (HEDI) and low (LEDI) eating disorder inventory scores.

 
Psychosocial Measures
The HEDI group, compared with the LEDI group, reported greater depressive symptoms on the Beck Depression Inventory (7.9 vs. 3.9, respectively; F(1,51) = 7.67, p < .01), greater trait anxiety on the Spielberger Trait Anxiety Inventory (46.8 vs. 34.2, respectively; F(1,49) = 27.14, p < .001), lower self-esteem (3.1 vs. 3.5, respectively; F(1,51) = 16.14, p < .001), and lower sense of mastery (3.1 vs. 3.3, respectively; F(1,51) = 5.17, p < .05). In terms of the Interpersonal Support Evaluation List, the HEDI group reported less appraisal support (17.1 vs. 19.7, respectively; F(1,51) = 9.58, p < .01), less tangible support (21.7 vs. 24.3, respectively; F(1,51) = 6.19, p < .05), and less social support in general (70 vs. 76.8, respectively; F(1,51) = 7.43, p < .01). There were no group differences, however, in satisfaction with social support as assessed with the Sarason Brief Social Support Questionnaire. The Profile of Mood States revealed that the HEDI group was less agreeable/more hostile (26.8 vs. 29.2, respectively; F(1,50) = 4.14, p < .05), less elated/more depressed (23.0 vs. 26.1, respectively; F(1,50) = 4.11, p < .05), less clearheaded/more confused (23.0 vs. 26.9, respectively; F(1,50) = 5.71, p < .05), less composed/more anxious (22.7 vs. 28.3, respectively; F(1,50) = 14.03, p < .001), and less confident/more unsure (19.1 vs. 24.4, respectively; F(1,50) = 10.74, p < .01). Although the groups did not significantly differ in perceived stress or in the number of daily stressors experienced, the Daily Stress Inventory did reveal that the emotional impact of stressors was greater in HEDI women (3.1 vs. 2.5, respectively; F(1,50) = 6.46, p = .01). Results from the Ways of Coping Scale revealed that the HEDI group blamed themselves significantly more (2.7 vs. 1.3, respectively; F(1,51) = 9.45, p < .01) and avoided their problems more (2.0 vs. 1.1, respectively; F(1,51) = 7.17, p = .01) when faced with stressors, but groups did not differ on the coping subscales of wishful thinking, seeking social support, minimizing threat, blaming others, counting blessings, and being problem focused.

24-Hour Ambulatory Mood and Blood Pressure Assessment
Mean Ambulatory Blood Pressure and Diary Measures.
Groups did not differ in overall 24-hour SBP, DBP, or MAP, nor did they differ in SBP, DBP, or MAP when examined during separate components of the 24-hour period, including day, evening, sleep, or morning BP. There were also no significant group differences in any diary measure.

The Effects of Mood, Tension, Stress, Urge to Eat, Hunger, and Satisfaction with Interpersonal Interactions on BP levels.
There was a main effect of negative/positive mood on SBP (F(2,39) = 4.5, p < .05), DBP (F(2,39) = 9.2, p = .004), and MAP (F(2,39) = 8.1, p < .01) because negative mood states were associated with greater blood pressure levels compared with positive mood states for all women. There was no Mood x Group interaction. There was also a main effect of relax/tense ratings on SBP (F(2,39) = 6.5, p = .01), DBP (F(2,39) = 7.3, p < .01), and MAP (F(2,39) = 8.1, p < .01) because instances of tension were associated with greater blood pressure levels compared with instances of relaxation for all women. There were no significant Tension x Group interactions. There were no effects involving urge to eat, satisfaction with interaction, mental activity, or hunger level for any measure of BP in either group.

24-Hour Urinary Neuroendocrine Measures.
The HEDI group had significantly lower 24-hour urinary NE levels compared with the LEDI group (F(1,48) = 4.8, p = .03) as depicted in Figure 2. There were no group differences in 24-hour EPI or DA excretions. On the other hand, the HEDI group had significantly greater 24-hour urinary cortisol levels compared with the LEDI group (F(1,49) = 7.33) as shown in Figure 3.



View larger version (10K):
[in this window]
[in a new window]
  		Fig. 2. Twenty-four hour NE (Mean + SEM) in women with high (HEDI) and low (LEDI) eating disorder inventory scores.

 


View larger version (9K):
[in this window]
[in a new window]
  		Fig. 3. Twenty-four hour urinary cortisol levels (Mean + SEM) in women with high (HEDI) and low (LEDI) eating disorder inventory scores.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
NOTES
 ACKNOWLEDGMENTS
 REFERENCES
 
The results of our study revealed that women who scored high on the EDI bulimia subscale but who had never met criteria for an eating disorder showed evidence of neuroendocrine dysregulation in SAM and HPAC axes, enhanced cardiovascular reactivity to a laboratory stressor, and differences in psychosocial profiles compared with women scoring lowest on the EDI bulimia subscale. Specifically, consistent with our a priori hypotheses, the HEDI women exhibited decreased urinary NE, increased urinary cortisol, and a more negative psychosocial profile than the LEDI group. Contrary to our expectations, however, HEDI women exhibited exaggerated, and not blunted, cardiovascular reactivity to laboratory stress relative to LEDI women, and there were no differences in 24-hour ambulatory blood pressure measures.

Regarding cardiovascular reactivity to the controlled, laboratory stressor, we found that the HEDI women had increased blood pressure and HR reactivity compared with women who scored lowest on the same subscale. It should be noted that our results for increased cardiovascular reactivity in HEDI women are in contrast with our previous findings in women with established BN, who showed blunted cardiovascular reactivity relative to controls (21). It is not clear why the HEDI women would differ in direction of cardiovascular response relative to those with established BN, but several possibilities exist. First, the HEDI and LEDI groups differed significantly in ethnic makeup, with the HEDI group containing fewer African Americans. Although only a few BN studies have included a sufficient number of nonwhites (46), those that have included minorities have found that BN is underrepresented in minority groups (1, 3, 7, 46, 47). For example, high school- and college-based studies provide evidence that only 1.2% to 4% of African Americans meet criteria for BN compared with 14% of whites (46, 47). In our previous study (21), only 7% of the women meeting DSM-IV criteria for BN were African American. Thus, despite the extra recruitment efforts that were taken to screen African American women in the present study (28% screened were African American), the final ethnic distribution into the HEDI and LEDI groups was consistent with ethnic differences in the prevalence of BN. Moreover, if ethnicity influenced CV reactivity in our sample, one would expect greater, and not lesser, BP reactivity in the LEDI group because studies indicate greater BP responses to stress in African Americans vs. whites (45).

A second possibility for the greater reactivity in HEDI women is that our LEDI women represent a "super normal" control group and that their significantly lower BP and HR reactivity to stress reflects abnormally low scores on the EDI inventory. Because there are no normative data available from U.S. populations using untransformed EDI scores (28), this possibility cannot be ruled out. To address this issue, we compared the psychosocial profiles of the LEDI women in the current study with healthy women of similar age who were tested in a previous study (21) and who were recruited from the population at large. There seemed to be no discernible difference between the LEDI women in the present study and healthy women from that previous study in any psychosocial measure assessed, including depression, anxiety, perceived stress, social support, and coping skills. Thus, the possibility that the observed differences in cardiovascular reactivity between the HEDI and LEDI groups in the present study reflects an abnormal psychological makeup of the LEDI women seems unlikely.

Furthermore, the present results are also in contrast to the study by Cattanach et al. (11), which found no difference in cardiovascular reactivity to psychological stress between groups who scored high vs. low on the same eating disorder subscale. It should be noted, however, that there are several methodological differences between our study and that of Cattanach et al. (11), and these may account for the discrepant findings. These include different criteria used to define the high vs. low EDI groups, lack of confirmation in the Cattanach et al. (11) study of whether any of their subjects met criteria for an eating disorder or other Axis I disorders, and the Cattanach et al. (11) study included only 15 eating disordered individuals, which may have decreased their power to detect significant effects.

In addition to our observation for greater cardiovascular reactivity in the HEDI women in the present study, we also observed that they had lower 24-hour urinary NE levels. This relationship between cardiovascular reactivity and NE may at first seem paradoxical because both are thought to reflect sympathetic nervous system (SNS) activation. However, consistent with pharmacological theory, catecholamine-induced changes in ß-adrenergic receptor responsivity may underlie this relationship. For example, the decreased 24-hour urinary NE evident in our HEDI group is consistent with studies by George et al. (48) and Pirke et al. (22), which found decreased basal plasma NE in women with established BN relative to controls. Of particular relevance is that George et al. (48) also found increased cardiac response to isoproterenol (ß-receptor agonist) in BN women vs. controls, suggesting that lower NE levels may be associated with upregulation of the cardiac ß-adrenergic receptors in bulimic women. If this is the case, this could account for both the increased cardiovascular reactivity and decreased 24-hour urinary NE level observed in the HEDI women in the present study.

In contrast to the urinary NE results, we observed increased 24-hour urinary cortisol levels in the HEDI women relative to the LEDI women. Although there have been reports of no differences in plasma cortisol between bulimic women and controls (49–52), there also exist a number of studies showing elevated basal plasma cortisol levels (22, 51, 53, 54), and our own prior study found increased cortisol in response to an ischemic pain task in bulimic women vs. controls (20). Additional evidence for hypercortisolemia in BN comes from studies using the dexamethasone suppression test, which found that bulimic women failed to show suppression of cortisol when challenged with an exogenous glucocorticoid (55–58). Thus, although there are discrepancies, the available evidence suggests that where differences in HPAC activation exist between bulimic women and controls, bulimic women show heightened HPAC activation. The present results extend the existing literature by demonstrating evidence for blunted NE plus increased HPAC activation in women with only eating disorder tendencies, but who have never met criteria for an eating disorder.

We also observed a psychosocial profile in these HEDI women that is very consistent with that seen in women with established BN. For example, HEDI women reported greater depressive symptoms, greater trait anxiety, lower self-esteem, lower sense of mastery, and less social support. The HEDI group was also more hostile, engaged in more negative coping strategies, and experienced a greater impact of stress than the LEDI women. Our findings for blunted NE output, combined with elevated cortisol output and negative psychosocial profiles in the HEDI women, is consistent with Frankenhaeuser’s model of effort vs. distress. Specifically, Frankenhaeuser’s model suggests that situations involving effort and positive affect (ie, effort) are associated with increases in catecholamine levels (59–64) but with low or even decreased cortisol secretion (60, 61, 63). On the other hand, a situation where negative affect is predominant and little effort is evident (ie, distress), a substantial increase in cortisol relative to catecholamine levels is exhibited (59–61, 63). Thus, the psychosocial and neuroendocrine profiles of our HEDI women are consistent with the distress scenario of Frankenhaeuser’s model.

Our study was the first to investigate the relationship between mood and ambulatory blood pressure in relation to eating disorder symptoms. Contrary to our expectations, however, although negative mood and increased tension were associated with greater BP levels in all women and although group differences in BP and HR appeared in response to a controlled laboratory stressor, there were no differences between HEDI and LEDI women in terms of overall 24-hour, day, evening, sleep, or morning BP. This is somewhat surprising as one might have expected that BP differences would at least emerge during a time period (ie, day) when the HEDI women were likely to experience an increase in perceived stress, whether due to negative mood and/or negative interpersonal interactions. One reason for the lack of ambulatory BP differences between groups may be that there were not enough instances of negative mood, interpersonal interactions, or stressful situations during the 24-hour monitoring period to detect an effect. Thus, a lack of statistical power may have contributed to the lack of a group difference in ambulatory BP. Future studies in larger samples using several days of ambulatory monitoring may be needed to detect differential relationships involving mood, stress, and BP in women with and without eating disorder tendencies.

In conclusion, we found that women with eating disorder tendencies but who had never met criteria for an eating disorder exhibited greater BP and HR reactivity to a controlled stressor, increased 24-hour urinary cortisol, decreased 24-hour urinary NE, and more negative psychosocial profiles compared with the control women. The neuroendocrine and psychosocial pattern is consistent with prior observations in women with established BN. One possibility is that this pattern of neuroendocrine dysregulation and negative psychosocial factors precedes the development of eating disorders, particularly BN. If this is the case, this would have implications for both treatment and early intervention. However, another possibility is that these neuroendocrine and psychosocial differences characterize or are a consequence of women at subthreshold for BN, who engage in some bulimic behaviors but who will not necessarily develop full-blown BN. However, during the SCID diagnostic interview, <10% of the HEDI women reported bingeing, none reported purging, and only one reported that she engaged in diet or exercise to control weight. The most frequently reported symptom was feeling loss of control over eating (19% of the HEDI women). Thus, it does not seem likely that sub-threshold levels of bingeing, purging, or excessive diet or exercise contributed directly to the observed group differences in neuroendocrine or CV measures. Clearly, future longitudinal studies will be necessary to assess whether the neuroendocrine, CV, or psychosocial disturbances observed in the HEDI women in the present study precede the development of an eating disorder. The results of our study are intended to be of heuristic value in this regard.


   ACKNOWLEDGMENTS
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 NOTES
 ACKNOWLEDGMENTS
REFERENCES
 
This research was supported by National Institutes of Health Grants GCRC-RR00046 and MHCRC-MH33127 and by the Foundation of Hope for Research and Treatment of Mental Illness.


   NOTES
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 NOTES
ACKNOWLEDGMENTS
 REFERENCES
 
Because ethnic status may influence cardiovascular indices and/or catecholamine levels (43–45), ethnic influence on group differences in reactivity to speech stress (see Results) seen in the HEDI vs. LEDI women was explored. Specifically, the 12 LEDI whites were compared with the 13 LEDI African Americans for differences in cardiovascular reactivity using ANOVA. There were no significant ethnic differences in SBP (p = .19), DBP (p = .22), MAP (p = .65), or HR (p = .76) reactivity to the speech stressor in this young, healthy female sample. Back

Received for publication February 17, 1999.

Revision received December 13, 1999.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 NOTES
 ACKNOWLEDGMENTS
 REFERENCES
 

  1. Bushnell JA, Wells JE, Hornblow AR, Oakley-Browne MA, Joyce P. Prevalence of three bulimia syndromes in the general population. Psychol Med 1990; 20: 671–80.[Medline]
  2. DSM-IV. Diagnostic and statistical manual of mental disorders. 4th ed. Washington DC: American Psychiatric Association; 1994.
  3. Fairburn CG, Beglin SJ. Studies of the epidemiology of bulimia nervosa. Am J Psychiatry 1990; 147: 401–08.[Abstract/Free Full Text]
  4. Garfinkel PE, Lin E, Goering P, Spegg C, Goldbloom DS, Kennedy S, Kaplan AA, Woodside DB. Bulimia nervosa in a Canadian community sample: prevalence and comparison of subgroups. Am J Psychiatry 1995; 152: 1952–8.
  5. Kendler KS, MacLean C, Neale M, Kessler R, Heath A, Eaves L. The genetic epidemiology of bulimia nervosa. Am J Psychiatry 1991; 148: 1627–37.[Abstract/Free Full Text]
  6. Striegel-Moore RH, Silberstein LR, Frensch P, Rodin J. A prospective study of disordered eating among college students. Int J Eat Disord 1989; 8: 499–509.
  7. Nevo S. Bulimic symptoms: prevalence and ethnic differences among college women. Int J Eat Disord 1985; 4: 151–68.
  8. Grissett NI, Norvell NK. Perceived social support, social skills, and quality of relationships in bulimic women. J Consult Clin Psychol 1992; 60: 293–9.[Medline]
  9. Lacey JH, Coker S, Birtchnell SA. Bulimia: factors associated with its etiology and maintenance. Int J Eat Disord 1986; 5: 475–87.
  10. Troop NA, Holbrey A, Trowler R, Treasure JL. Ways of coping in women with eating disorders. J Nerv Ment Dis 1994; 182: 535–40.[Medline]
  11. Cattanach LM, Malley R, Rodin J. Psychologic and physiologic reactivity to stressors in eating disordered individuals. Psychosom Med 1988; 50: 591–9.[Abstract/Free Full Text]
  12. Johnson C, Berndt DJ. Preliminary investigation of bulimia and life adjustment. Am J Psychiatry 1983; 140: 774–7.[Abstract/Free Full Text]
  13. Kent JS, Clopton JR. Bulimia: a comparison of psychological adjustment and familial characteristics in a nonclinical sample. J Clin Psychol 1988; 44: 964–71.[Medline]
  14. Ordman AM, Kirschenbaum DS. Bulimia: assessment of eating, psychological adjustment, and familial characteristics. Int J Eat Disord 1986; 5: 865–78.
  15. Soukup VM, Beiler ME, Terrell F. Stress, coping style, and problem solving activity among eating-disordered inpatients. J Clin Psychol 1990; 46: 592–9.[Medline]
  16. Grace PS, Jacobson RS, Fullager CJ. A pilot comparison of purging and nonpurging bulimics. J Clin Psychol 1985; 41: 173–80.[Medline]
  17. Leon GR, Carroll K, Chernyk B, Finn S. Binge eating and associated habit patterns within college students and identified bulimic populations. Int J Eat Disord 1985; 4: 43–57.
  18. Pyle RL, Mitchell JE, Eckert ED. Bulimia: a report of 34 cases. J Clin Psychiatry 1981; 42: 60–4.[Medline]
  19. Cattanach LM, Rodin J. Psychosocial components of the stress process in bulimia. Int J Eat Disord 1988; 7: 75–88.
  20. Girdler SS, Koo-Loeb JH, Pedersen CA, Brown HJ, Maixner W. Blood pressure-related hypoalgesia in bulimia nervosa. Psychosom Med 1998; 60: 736–43.[Abstract/Free Full Text]
  21. Koo-Loeb JH, Pedersen CA, Girdler SS. Blunted cardiovascular and catecholamine stress reactivity in women with bulimia nervosa. Psychiatry Res 1998; 80: 13–27.[Medline]
  22. Pirke KM, Platte P, Laessle R, Seidl M, Fichter MM. The effect of a mental challenge test of plasma norepinephrine and cortisol in bulimia nervosa and in controls. Biol Psychiatry 1992; 32: 202–6.[Medline]
  23. Obrist PA. The cardiovascular-behavioral interaction: as it appears today. Psychophysiology 1976; 13: 95–107.[Medline]
  24. Peckerman A, Saab PG, McCabe PM, Skyler JS, Winters RW, Llabre MM, Schneiderman N. Blood pressure reactivity and the perception of pain during the forehead cold pressor test. Psychophysiology 1991; 28: 485–95.[Medline]
  25. Garner DM, Olmsted MP, Polivy J. Development and validation of a multidimensional eating disorder inventory for anorexia nervosa and bulimia. Int J Eat Disord 1983; 2: 15–34.
  26. Wear RW, Pratz O. Test-retest reliability for the Eating Disorder Inventory. Int J Eat Disord 1987; 6: 767–9.
  27. Raciti MC, Norcross JC. The EAT and EDI: screening, interrelationships, and psychometrics. Int J Eat Disord 1987; 6: 579–86.
  28. Schoemaker C, Van Strien T, Van Der Staak C. Validation of the eating disorders inventory in a nonclinical population using transformed and untransformed responses. Int J Eat Disord 1994; 15: 387–93.[Medline]
  29. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. An inventory for measuring depression. Arch Gen Psychiatry 1961; 4: 561–71.
  30. Spielberger CD, Gorusch RL, Lushene RE. STAI manual for the State-Trait Anxiety Inventory. Palo Alto: Consulting Psychologists Press; 1970.
  31. Rosenberg M. Society and the adolescent self-image. Princeton NJ: Princeton University Press; 1965.
  32. Pearlin LI, Menaghan EG, Lieberman MA, Mullan JT. The stress process. J Health Soc Behav 1981; 22: 337–56.[Medline]
  33. Cohen S, Mermelstein R, Kamarck T, Hoberman HM. Measuring, the functional components of social support. In: Sarason IG, Sarason BR, editors. Social support: theory, research and applications. Dordrecht: Martinus Nijhoff; 1985. p. 73–94.
  34. Lorr M, McNair DM. Profile of mood states: bipolar form. San Diego: Educational and Industrial Testing Service; 1988.
  35. Cohen S, Kamarck T, Mermelstein R. A global measure of perceived stress. J Health Soc Behav 1983; 24: 385–96.[Medline]
  36. Sarason IG, Sarason BR, Shearin EN, Pierce GR. A brief measure of social support: practical and theoretical implications. J Soc Person Relationships 1987; 4: 497–510.
  37. Lazarus RS, Folkman S. Stress, appraisal and coping. New York: Springer Publishing Company; 1984.
  38. Brantley PJ, Waggoner CD, Jones GN, Rappaport NB. A daily stress inventory: development, reliability, and validity. J Behav Med 1987; 10: 61–74.[Medline]
  39. Ainsworth BE, Haskell WL, Leon AS, Jacobs DR, Montoye HJ, Sallis JF, Paffenbarger RS. Compendium of physical activities: classification of energy costs of human physical activities. Med Sci Sports Exerc 1993; 25: 71–80.[Medline]
  40. Light KC, Obrist PA, Cubeddu LX. Evaluation of a new ambulatory blood pressure monitor (Accutracker 102): laboratory comparisons with direct arterial pressure, stethoscopic auscultatory pressure, and readings from a similar monitor (Spacelabs Model 5200). Psychophysiology 1988; 25: 107–16.[Medline]
  41. Girdler SS, Turner JR, Sherwood A, Light KC. Gender differences in blood pressure control during a variety of behavioral stressors. Psychosom Med 1990; 52: 571–91.[Abstract/Free Full Text]
  42. Saab PG, Matthews KA, Stoney CM, McDonald RH. Premenopausal and postmenopausal women differ in their cardiovascular and neuroendocrine responses to behavioral stressors. Psychophysiology 1989; 26: 270–80.[Medline]
  43. Ironson GH, Gellman MD, Spitzer SB, Llabre MM, De Carlo Pasin R, Weidler DJ, Schneiderman N. Predicting home and work blood pressure measurements from resting baseline and laboratory reactivity in black and white Americans. Psychophysiology 1989; 26: 174–84.[Medline]
  44. Light KC, Turner JR, Hinderliter AL, Girdler SS, Sherwood A. Comparison of cardiac versus vascular reactors and ethnic groups in plasma epinephrine and norepinephrine responses to stress. Int J Behav Med 1994; 1: 229–46.[Medline]
  45. Light KC, Turner JR, Hinderliter AL, Sherwood A. Race and gender comparisons: I. hemodynamic responses to a series of stressors. Health Psychol 1993; 12: 354–65.[Medline]
  46. Pike KM, Walsh BT. Ethnicity and eating disorders: implications for incidence and treatment. Psychopharmacol Bull 1996; 32: 265–74.[Medline]
  47. Gray JJ, Ford K, Kelly LM. The prevalence of bulimia in a black college populations. Int J Eat Disord 1987; 6: 733–40.
  48. George DT, Kaye WH, Goldstein DS, Brewerton TM, Jimerson DC. Altered norepinephrine regulation in bulimia: effects of pharmacological challenge with Isoproterenol. Psychiatry Res 1990; 33: 1–10.[Medline]
  49. Gwirtsman HE, Kaye WH, George DT, Jimerson DC, Ebert MH, Gold PW. Central and peripheral ACTH and cortisol levels in anorexia nervosa and bulimia. Arch Gen Psychiatry 1989; 46: 61–9.[Abstract/Free Full Text]
  50. Hudson JI, Katz DL, Pope HG, Hudson MS, Griffing GT, Melby JC. Urinary free cortisol and response to the dexamethasone suppression test in bulimia: a pilot study. Int J Eat Disord 1987; 6: 191–8.
  51. Laue L, Gold PW, Richmond A, Chrousos GP. The hypothalamic-pituitary-adrenal axis in anorexia nervosa and bulimia nervosa: pathophysiologic implications. Adv Pediatr 1991; 38: 287–316.[Medline]
  52. Newman MM, Halmi KA. The endocrinology of anorexia nervosa and bulimia nervosa. Neurol Clin 1988; 6: 195–212.[Medline]
  53. Kennedy SH, Garfinkel PE, Parienti V, Costa D, Brown GM. Changes in melatonin levels but not cortisol levels are associated with depression in patients with eating disorders. Arch Gen Psychiatry 1989; 46: 73–8.[Abstract/Free Full Text]
  54. Mortola JF, Rasmussen DD, Yen SS. Alterations of the adrenocorticotropin-cortisol axis in normal weight bulimic women: evidence for a central mechanism. J Clin Endocrinol Metab 1989; 68: 517–22.[Abstract/Free Full Text]
  55. Gwirtsman HE, Roy-Byrne P, Yager J, Gerner RH. Neuroendocrine abnormalities in bulimia. Am J Psychiatry 1983; 140: 559–63.[Abstract/Free Full Text]
  56. Hudson JI, Pope HG, Jones JM, Yurgelun-Todd D. Hypothalamic-pituitary-adrenal axis hyperactivity in bulimia. Psychiatry Res 1983; 8: 111–8.[Medline]
  57. Lindy DC, Walsh BT, Roose SP, Gladis M, Glassman AH. The dexamethasone suppression test in bulimia. Am J Psychiatry 1985; 142: 1375–6.[Abstract/Free Full Text]
  58. Mitchell JE, Pyle RL, Hatsukami D, Boutzcoff LI. The dexamethasone suppression test in patients with bulimia. J Clin Psychiatry 1985; 45: 508–11.
  59. Frankenhaeuser M. Psychoneuroendocrine approaches to the study of stressful person-environment transactions. In: Selye H, editor. Guide to stress research. Vol I. New York: Nostrand Reinhold; 1980. p. 46–70.
  60. Frankenhaeuser M. The sympathetic-adrenal and pituitary-adrenal response to challenge: comparison between the sexes. In: Dembroski TM, Schmidt TH, Blumcher G, editors. Biobehavioral bases of coronary heart disease. Basel: Kanger; 1983. p. 91–105.
  61. Frankenhaeuser M. A biopsychosocial approach to work life issues. Int J Health Serv 1989; 19: 747–58.[Medline]
  62. Frankenhaeuser M, Lundberg U, Fredrikson M, Melin B, Tuomisto M, Myrsten A, Hedman M, Bergman-Losman B, Wallin L. Stress on and off the job as related to sex and occupational status in white-collar workers. J Organizational Behav 1989; 10: 321–46.
  63. Lundberg U, Frankenhaeuser M. Pituitary-adrenal and sympathetic-adrenal correlates of distress and effort. J Psychosom Res 1980; 24: 125–30.[Medline]
  64. Pollard TM, Ungpakorn G, Harrison GA, Parkes KR. Epinephrine and cortisol responses to work: a test of the models of Frankenhaeuser and Karasek. Ann Behav Med 1996; 18: 229–37[Medline]



This article has been cited by other articles:


Home page
 Psychosom. Med.Home page
B. Ludescher, G. Leitlein, J.-E. Schaefer, S. Vanhoeffen, S. Baar, J. Machann, C. D. Claussen, F. Schick, and G. W. Eschweiler
Changes of Body Composition in Bulimia Nervosa: Increased Visceral Fat and Adrenal Gland Size
Psychosom Med, January 1, 2009; 71(1): 93 - 97.
[Abstract] [Full Text] [PDF]

Home page
 Psychosom. Med.Home page
M. E. Gluck, A. Geliebter, J. Hung, and E. Yahav
Cortisol, Hunger, and Desire to Binge Eat Following a Cold Stress Test in Obese Women With Binge Eating Disorder
Psychosom Med, November 1, 2004; 66(6): 876 - 881.
[Abstract] [Full Text] [PDF]

Home page
 Arch Gen PsychiatryHome page
J. G. Johnson, P. Cohen, S. Kasen, and J. S. Brook
Eating Disorders During Adolescence and the Risk for Physical and Mental Disorders During Early Adulthood
Arch Gen Psychiatry, June 1, 2002; 59(6): 545 - 552.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowRequest Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Koo-Loeb, J. H.
Right arrow Articles by Girdler, S. S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Koo-Loeb, J. H.
Right arrow Articles by Girdler, S. S.
Related Collections
Right arrow Neuroendocrine
Right arrow Eating Disorder

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS