| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Center for Environmental Hazards Research (A.P., J.J.L., S.L.S., A.T., L.T., C.P., L.R.K., J.E.O., B.H.N.), VA Medical Center, East Orange, NJ; Environmental and Occupational Health Sciences Institute (A.P.), Robert Wood Johnson Medical School, Piscataway, NJ; and Department of Psychology (B.H.), University of Miami, Coral Gables, FL.
Address reprint requests to: Arnold Peckerman, PhD, Gulf War Research Center (127B), VA Medical Center, East Orange, NJ 07018-1095. Email: apeckerm{at}nbunj.jvnc.net
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
METHODS: Psychophysiological stress testing was performed on 51 Gulf War veterans with chronic fatigue (using the 1994 case definition of the Centers for Disease Control and Prevention) and 42 healthy veterans. Hemodynamic responses to cold pressor, speech, and arithmetic stressors were evaluated using impedance cardiography.
RESULTS: Veterans with chronic fatigue had diminished blood pressure responses during cognitive (speech and arithmetic) stress tests due to unusually small increases in total peripheral resistance. The cold pressor test, however, evoked similar blood pressure responses in the chronic fatigue and control groups. Low reactivity to cognitive stressors was associated with greater fatigue ratings among ill veterans, whereas an opposite relation was observed among healthy veterans. Self-reported neurocognitive decline was associated with low reactivity to the arithmetic task.
CONCLUSIONS: These results suggest a physiological basis for some Gulf War veterans’ reports of severe chronic fatigue. A greater deficit with responses processed through cerebral centers, as compared with a sensory stimulus (cold pressor), suggests a defect in cortical control of cardiovascular function. More research is needed to determine the specific mechanisms through which the dissociation between behavioral and cardiovascular activities identified in this study may be contributing to symptoms in Gulf War veterans.
Key Words: Gulf War illnesses • chronicfatigue • cardiovascular stress response • impedance cardiography
Abbreviations: AD ACL = Activation-Deactivation Adjective Checklist; BDI= Beck Depression Inventory; CDC = Centers for Disease Control andPrevention; CF = chronic fatigue; CFS = chronic fatiguesyndrome; DBP = diastolic blood pressure; GV = Gulf Warveteran; HR = heart rate; ICF = idiopathic chronic fatigue; MAP = mean arterial pressure; MFI = Multidimensional FatigueInventory; PEP = preejection period; POMS = Profile of MoodStates; 
= cardiac output; SBP = systolic bloodpressure; SV = stroke volume; TPR = total peripheral resistance.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
The study reported here examined the possibility that a problem with circulation may underlie symptoms in GVs meeting the case definition of CF (9). Our basic hypothesis was that GVs with CF have abnormalities in central regulatory mechanisms of blood pressure control that might result in an inadequate perfusion. To test this hypothesis, we examined whether hemodynamic responses to behavioral stressors were altered in GVs with CF. The study was designed to provide preliminary data on what types of challenges (cognitive vs. physical) and specific mechanisms (ie, those concerned with regulation of peripheral resistance vs. those regulating cardiac output) may be involved. In addition, the study examined whether changes in cardiovascular functioning in GVs with CF were consistent with major symptoms of this illness, including fatigue, cognitive abnormalities, and associated functional decline. With respect to fatigue, our basic presumption was that lack of energy in ill veterans was caused by a disease and thus would have a different physiological basis compared with fatigue caused by overexertion or other transient factors. Also examined were the effects of depressive and anxious symptomology on stress responses in GVs with CF. Although comorbid psychiatric disorders are common in ill GVs, their relation to physical symptoms remains unclear (3, 10).
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
Control subjects were GVs in good health, which was verified by on-site medical evaluation. None of the subjects reported a history of cardiovascular, respiratory, or neurological disorders or used medications with central or peripheral adrenergic activity. The CF and control groups had similar composition with respect to gender (men/women ratio = 2.6) and race (78% white, 19% African American, and 3% Asian) and were similar in age (35 ± 8 years), body mass index (27 ± 4 kg/m2), alcohol consumption (3 ± 6 drinks/wk), and caffeine consumption (216 ± 210 mg/d). The CF group had more years of education than the control group (16 ± 2 vs. 14 ± 2, p < .0005) and a greater percentage of smokers (25% vs. 7%, p < .05). Cigarette consumption per smoker was not significantly different in the CF and control groups (14 ± 8 cigarettes/d).
As part of the initial health survey, subject completed the MFI (13) and the Karnofsky Illness Severity Scale (14). The MFI is a validated 20-item questionnaire that yields scores on general, physical, and mental fatigue, reduced activity, and reduced motivation. For purposes of this study, responses on the MFI were scored on the subscale of reduced activity, which provides a measure of functional status during the previous month. The Karnofsky scale provides a measure of global disability in terms of overall physical impairment during the past week. GVs with CF had reduced functionality on both the MFI reduced activity scale (13 ± 4 vs. 6 ± 3 in control subjects) and the Karnofsky global disability scale (60 ± 19% vs. 98 ± 5% in control subjects). GVs with CF also had more symptoms of depression and anxiety than control subjects, as assessed by the Beck BDI (15) and the POMS tension-anxiety subscale (16) (16 ± 10 vs. 2 ± 3 and 13 ± 8 vs. 5 ± 4, respectively; p values < .0001).
Study Design and Procedures
All subjects were required to abstain from nicotine for 2 hours and from caffeine for 4 hours before laboratory testing. The experimenter in contact with the subject was not informed1 of the veteran’s clinical status. Standard information about the study was provided at the beginning of the laboratory session. The subject was told that the purpose of the study was to examine functioning of the cardiovascular system in GVs. Before laboratory testing began, the subject completed the AD ACL (17), a 20-item scale that yields bipolar dimensions of energetic and tense arousal states. Scores on the energetic arousal dimension (energy scores) were used to estimate severity of fatigue on the day of testing. All subjects were tested in the morning. The complete protocol consisted of standard autonomic testing (18), which was done first over a 40-minute period, followed by behavioral stress testing. The results of autonomic testing will be reported elsewhere. The behavioral stress battery included the forehead cold pressor test, speech presentation, and a mental arithmetic task, all performed with the subject seated. The cold pressor test was given first, and speech and arithmetic tasks were given next in a counterbalanced order. A 20-minute rest period preceded the cold pressor test and the speech-arithmetic series of stressors. The cold pressor test consisted of applying a plastic bag containing a mixture of crushed ice and water (1°C) to the forehead for 2 minutes and included ratings of pain on a scale of 0 to 10 at 5-second intervals (19). The forehead cold pressor test was chosen over more traditional versions of the test because it does not evoke an increase in HR (19), with changes in SV being more reflective of left ventricular performance under conditions of increased afterload. For the speech task, the subject was given a scenario involving being accused of shoplifting and was given 3 minutes to prepare and 3 minutes to deliver a speech about the situation into a video camera (20). The arithmetic stressor was 3 minutes in duration and consisted of subtractions carried out on pairs of two- to three-digit numbers presented to the subject at 10-second intervals by means of an audiotape. The task was performed under conditions of audiovisual interference, which consisted of a TV playing a prerecorded daytime talk show in front of the subject at the same loudness level as the audiotape (about 70 dB). Ratings of perceived stress, challenge, and threat were obtained for each baseline and stressor period using a Likert scale with ratings of 1 to 8.
Physiological Measurements
Blood pressure was recorded in the left arm using a Dinamap 1846 SX automated blood pressure monitor. Impedance cardiography data were recorded with a University of Miami impedance cardiograph (model R03) using previously described methods (19). The data were scored with an automated software system (21, 22) that provides values for HR, SV, , TPR, and PEP. The peak amplitude of the impedance cardiogram was measured relative to the isoelectric line, as recommended by Sherwood et al. (23). Reproducibility of impedance cardiography measurements during the behavioral tasks has been demonstrated (24). Two 30-second samples were collected during the last 2 minutes of rest periods and during the last 1.5 minutes of stressor periods and were averaged to represent the baseline and test response levels.
Data Analysis
Data were analyzed using SAS statistical software (SAS Institute, Cary, NC). The effects of illness on responses to stressors were examined in mixed analyses of variance. Two subjects in each group did not complete the cold pressor test because of pain, and three CF subjects declined to complete the speech task, resulting in missing data for seven subjects. Values for SV, , and TPR were normalized using log transformation before analyses. A Greenhouse-Geisser correction (25) was applied in all analyses of repeated measures. Multiple regression analysis (26) was used to examine the relationship between symptoms and cardiovascular stress responses. Because changes in and TPR are interdependent (27), regression parameters describing their relationship to independent variables were computed with the other variable in the model. Controlling for their shared variance (mean R2 = 0.43 in the present study) was aimed at improving the sensitivity of 
and TPR scores as indices of task-induced cardiac and vasomotor responses, respectively. The rationale for this procedure and its application has been described elsewhere (19). All data in the text are reported as mean ± SD. A p value < .05 was considered significant in all analyses except when testing for interactions in multiple regression analyses, in which a value of p < .10 was used (28).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
(p values < .05), and TPR (p < .003). Post hoc analyses of simple interactions revealed smaller increases in SBP and DBP in the CF group during the speech (p values < .0005) and arithmetic (p values < .05) tasks but not during the cold pressor test (Table 1). The CF group had smaller increases in TPR during all stressor periods (p values < .05, .005, and .05, for the cold pressor, speech, and arithmetic tasks, respectively); these increases in TPR were associated with a greater increase in during the speech task (simple interaction, p < .01) and a maintained (vs. declined in control subjects, p < .01) during the cold pressor periods. HR increased, SV declined, and PEP shortened during the speech and arithmetic tasks (p values < .05) with no significant effects of groups (p values > .10). At the end of the recovery period after the cold pressor test, SBP, DBP, and TPR remained elevated and SV was reduced relative to the values before this test (p values < .05), but there were no effects of groups (simple interactions, p values > .50).
|
Self-Perceived Energy and Cardiovascular Stress Responses
As measured on the day of testing, GVs with CF had lower AD ACL energy scores than control subjects (adjusted mean, 8 ± 6 vs. 17 ± 6; p < .0001). Therefore, one hypothesis that might have explained the hyporeactivity in GVs with CF was that it was simply due to their lower energy levels. However, another possibility was that abnormal stress responses and severe fatigue in the patient group were both expressions of the same pathological condition. To test these hypotheses, the next set of analyses explored the relationship between energy scores and cardiovascular responses to cognitive stressors using multiple regression analysis. Cardiovascular variables in these analyses were modeled as a function of AD ACL energy scores, CF diagnosis, and their interaction, controlling for differences in years of education and smoking status. Significant interaction between illness and energy levels in these analyses would be consistent with the latter but not the former hypothesis (ie, it would suggest that a pathological condition had altered the relationship between fatigue and reactivity in the CF group). The BDI depression and POMS anxiety scores were also included in the model to determine whether the psychological symptoms were a factor in hyporeactivity. All predictor variables were retained in the model regardless of their significance for control purposes. The dependent variables in these analyses were changes in blood pressure, , and TPR during cognitive stress tests. After preliminary analyses found similar results for the two cognitive tasks, the tests were performed using the mean speech-arithmetic change scores. Also, because SBP and DBP did not provide substantially different information, changes in MAP (mean speech-arithmetic values, 9 ± 7 vs. 14 ± 6 mm Hg in the CF and control groups, respectively) were used as measures of blood pressure response.
The results of these analyses are shown in Table 2. It can be seen that energy by group interactions were significant for all criterion variables. Post hoc analyses of simple regressions revealed that low energy scores predicted smaller increases in MAP, , and TPR in the CF group, whereas in the control group the relationship was nonsignificant for MAP and was actually negative for (Figure 1). Feeling anxious or depressed was unrztrelated to cardiovascular stress responses in these analyses (Table 2). Interestingly, having CF was associated with hyporeactivity even after accounting for the effects of low energy (Table 2).
|
|
, and TPR, while controlling for differences in arithmetic performance and energy scores, using multiple regression analysis. Supporting our hypothesis, self-rated cognitive impairment was a significant predictor of diminished increases in MAP during the arithmetic stressor but not during the speech stressor (Table 3). Significant associations were also present between cognitive impairment scores and reduced and TPR responses to the arithmetic task (Table 3). There was only a weak correlation between cognitive impairment scores and performance on the arithmetic test (r = -0.29, p < .04), and GVs with CF who rated their cognitive impairment as substantial or greater (N = 38) did not perform significantly worse on the arithmetic task than those who rated this problem as moderate or less severe (N = 13) (percentage correct, 53 ± 18% vs. 62 ± 14%; p > .14). However, maintained performance in this severely impaired group seemed to come at the cost of greater effort, as indicated by the higher ratings of perceived stressfulness of the arithmetic task (means adjusted for pretask level, 4.7 ± 2 vs. 3.2 ± 2; p < .009).
|
responses (adjusted for TPR, semipartial r = -0.24; p < .04). No other significant correlations between functional status and hemodynamic stress responses were observed.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
, with a similar trend seeming to affect changes in MAP and TPR. These findings suggest that fatigue is associated with reduced cardiovascular responsiveness to stressors when it is pathological (ie, unremitting and debilitating), whereas the opposite may be the norm when fatigue is "normal" (ie, sporadic and relieved by rest). The most conspicuous hemodynamic response deficit to emerge in veterans with CF was the failure of TPR to increase in response to challenges. This failure was the apparent cause of low blood pressure during cognitive stress tests but also occurred during the cold pressor test. Thus, GVs with CF had a problem controlling peripheral vascular resistance during mental stress, when the primary input is coming from cortical centers and is executed by increased sympathetic outflow to visceral arterial beds (29), as well as during stimulation of cutaneous cold receptors, which reflexively increase sympathetic vasoconstrictor nerve traffic to the skin (30).
Although no causal attributions are possible from the present data, there was a clear and robust relationship between cardiovascular function and illness in these veterans. Ill veterans with the most impaired stress responses were more fatigued and functionally impaired than those whose responses were closer to the norm. Furthermore, GVs who had a severe problem with memory or concentration had low blood pressure responses specifically during the mental arithmetic task, implying a direct relation between illness and cardiovascular function. Reduced blood pressure responses to mental arithmetic in these patients were associated with greater perceived stress during the task. In other words, the same amount of mental work seemed to require greater effort on their part. These data suggest that one consequence of reduced cardiovascular support of mental activities may be a greater required effort.
Overall, this study identifies a pattern of dissociations between behavioral and cardiovascular activities in GVs with CF that may be contributing to their symptoms by causing underperfusion of the brain and peripheral organs. This interpretation is consistent with reports that patients with CFS have reduced blood flow in the brain (8) and muscle (31). Because this study was conducted in a self-selected, and perhaps more motivated or otherwise unique, sample of patients, its results may not be generalizable to all GVs with CF. We did, however, find a similar pattern of alterations in cardiovascular stress responses in CF patients in the general population (32), suggesting that it may be a common condition in unexplained CF.
Low blood pressure responses to tasks requiring high-level central processing contrasted with normal responses to the cold pressor test, a sensory stressor with aversive characteristics. This pattern of results seemed to suggest a problem in the brain areas interfacing cognitive and autonomic activities. Several studies have found that an inability to regulate blood pressure during mental challenges can be a symptom of a brain disease (33, 34). Consistent with this hypothesis, GVs with symptoms resembling CF were reported to display neurological signs consistent with injury to brain areas involved in autonomic regulation (35).
Physiological responses to a stressor are shaped by the appraisal it receives and by the type of a behavioral response it elicits (36). Speech and arithmetic tasks require active engagement and concentrated mental effort. Research on cognitive determinants of cardiovascular stress responses (37–39) has shown that multiple aspects of the engagement with a task can have significant influence on the evoked cardiovascular responses. Therefore, although cognitive stressors used in the study were perceived as stressful, challenging, and threatening to a similar extent by case and control subjects, the possibility remains that fatigue-induced disengagement provided at least a partial explanation for group differences in cardiovascular stress responses.
The reduction in functional status reported by patients with CF was proportionately greater in those with the lowest blood pressure responses to cognitive stressors. A sedentary lifestyle associated with CF may cause a reduction in fitness level that might, in turn, affect cardiovascular responsiveness to stressors. It is unlikely, however, that cardiovascular hyporesponsiveness in GVs with CF can be explainable by deconditioning because sedentary individuals tend to be more, rather than less, reactive to behavioral stressors (40, 41). Rather, our data suggest that the opposite relation, such that inadequate physiological responses during cognitive activities further diminish the patient’s work capacity, could explain the association between low blood pressure responses and a decline in functional status.
A substantial proportion of GVs with illnesses have comorbid psychiatric conditions, most commonly depressive and anxiety disorders (3, 10). However, blunted stress responses are the opposite of the outcomes associated with depression and anxiety, in which a common finding is upregulation of sympathetic activity (42, 43) with increased responsiveness to behavioral stressors (44). Not surprisingly, therefore, symptoms of depression and anxiety did not explain hyporeactivity in multiple regression analyses. These results support a hypothesis that psychiatric symptoms in GVs are distress reactions to chronic illness rather than a primary disorder.
It must be noted that if the shared variance between changes in and TPR had not been controlled, we would have found fewer positive results. Specifically, the differential relationship between energy scores and TPR responses to cognitive stressors in the CF and control groups would not have been detected (interaction, p > .38), and the association between cognitive symptoms and depressed and TPR responses to the arithmetic task in veterans with CF would have been completely obscured (p values > .32). These results suggest that this statistical control procedure might be useful in clarifying the effector mechanisms of blood pressure responses to stressors. Additional experimental work is required to confirm the value of this statistical approach.
To conclude, our findings suggest that symptoms of illness in GVs with CF are linked to the circulation in a coherent and physiologically significant way. Additional research is needed to define the specific mechanisms of cardiovascular hyporesponsiveness in Gulf War–related fatiguing illness, most notably whether the problem with vasoconstriction is centrally or peripherally mediated. Other important questions that deserve more study include the extent to which a lack of engagement in the task is a significant factor in hyporeactivity in veterans with CF and how this may influence other aspects of physiological functioning and health.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
Inclusion of order (speech-arithmetic) and group by order terms in the mixed model revealed no differential carryover effects (all p values > .10) and did not affect the reported results.
Received for publication November 19, 1998.
Revision received February 1, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  R. Nelesen, Y. Dar, K. Thomas, and J. E. Dimsdale The Relationship Between Fatigue and Cardiac Functioning Arch Intern Med, May 12, 2008; 168(9): 943 - 949. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Fiedler, N. Giardino, B. Natelson, J. E. Ottenweller, C. Weisel, P. Lioy, P. Lehrer, P. Ohman-Strickland, K. Kelly-McNeil, and H. Kipen Responses to Controlled Diesel Vapor Exposure Among Chemically Sensitive Gulf War Veterans Psychosom Med, July 1, 2004; 66(4): 588 - 598. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Peckerman, J. J. LaManca, B. Qureishi, K. A. Dahl, R. Golfetti, Y. Yamamoto, and B. H. Natelson Baroreceptor Reflex and Integrative Stress Responses in Chronic Fatigue Syndrome Psychosom Med, September 1, 2003; 65(5): 889 - 895. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. J. LaManca, A. Peckerman, S. A. Sisto, J. DeLuca, S. Cook, and B. H. Natelson Cardiovascular Responses of Women With Chronic Fatigue Syndrome to Stressful Cognitive Testing Before and After Strenuous Exercise Psychosom Med, September 1, 2001; 63(5): 756 - 764. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
