Cardiovascular Responses Of Women With Chronic Fatigue Syndrome To Stressful

Cardiovascular Responses of Women With Chronic Fatigue Syndrome to Stressful Cognitive Testing Before and After Strenuous Exercise JOHN J. LAMANCA, PHD, ARNOLD PECKERMAN, PHD, SUE ANN SISTO, PT, PHD, JOHN DELUCA, PHD, SEAN COOK, BS, AND BENJAMIN H. NATELSON, MD Objective: The purpose of this study was to compare the cardiovascular responses of patients with chronic fatigue syndrome (CFS) to healthy control subjects when performing stressful cognitive tasks before and after strenuous exercise. Method: Beat-by-beat blood pressure and electrocardiogram were recorded on 19 women with CFS and 20 healthy nonexercising (ie, sedentary) women while they performed cognitive tests before, immediately after, and 24 hours after incremental exercise to exhaustion. Results: Diminished heart rate (p ⬍ .01) and systolic (p ⬍ .01) and diastolic (p ⬍ .01) blood pressure responses to stressful cognitive testing were seen in patients with CFS when compared with healthy, sedentary controls. This diminished stress response was seen consistently in patients with CFS across three separate cognitive testing sessions. Also, significant negative correlations between self-ratings of CFS symptom severity and cardiovascular responses were seen (r ⫽ ⫺0.62, p ⬍ .01). Conclusions: Women with CFS have a diminished cardiovascular response to cognitive stress; however, exercise did not magnify this effect. Also, the data showed that the patients with the lowest cardiovascular reactivity had the highest ratings of CFS symptom severity, which suggests that the individual response of the patient with CFS to stress plays a role in the common complaint of symptoms worsening after stress. Key words: chronic fatigue, stress response, cognitive performance, exercise.

CFS ⫽ chronic fatigue syndrome; HR ⫽ heart rate; SBP ⫽ systolic blood pressure; DBP ⫽ diastolic blood pressure; PP ⫽ pulse pressure; MAP ⫽ mean arterial blood pressure; SCWT ⫽ Stroop Color/Word Interference test; SDMT ⫽ Symbol Digit Modalities test; AD ACL ⫽ Activation-Deactivation Adjective Check List; ECG ⫽ electrocardiogram; BDI ⫽ Beck ˙ O2peak ⫽ peak oxygen conDepression Inventory; V sumption.

CFS is characterized by debilitating fatigue that cannot be attributed to any known medical cause and is associated with at least a 50% decrease in daily activity from premorbid levels (1). One major patient complaint is that CFS symptoms get substantially worse after physical or mental stress (1). This suggests that the symptoms of CFS may result from an abnormal physiological response to physical or mental challenges (2). This hypothesis is supported by the fact that a reduced stress response system is associated with prolonged fatigue and immunological changes, which

From the Chronic Fatigue Syndrome Cooperative Research Center, University of Medicine and Dentistry of New Jersey—New Jersey Medical School, Newark, New Jersey (J.J.L., S.A.S., J.D., S.C., B.H.N.); The Heart Failure Center, Division of Circulatory Physiology, New York Presbyterian Hospital New York, New York (J.J.L.); and Kessler Medical Rehabilitation Research and Education Corporation, West Orange, New Jersey (S.A.S., J.D.). Address reprint requests to: Benjamin H. Natelson, MD, Fatigue Research Center, New Jersey Medical School, 88 Ross Street, East Orange, NJ 07018. Received for publication September 16, 1999; revision received January 11, 2001.

756 0033-3174/01/6305-0756 Copyright © 2001 by the American Psychosomatic Society

have been thought to contribute to the pathophysiology of CFS (3, 4). Physical stress affects patients with CFS differently from healthy, sedentary controls. For example, we have reported that patients with CFS have a significant decrease in cognitive functioning (5) and daily ambulatory activity levels (6) after an exercise test to exhaustion. Also, we have seen lower vagal activity in patients with CFS when compared with controls during and after mild exercise (7). Others have reported cognitive deficits after exercise (8). One research theory seeking to explain the inability of the patient with CFS to respond normally during and after physical stress has to do with ineffective autonomic control of blood pressure. The hypothesis is that an inadequate blood pressure response to stress leads to cerebral hypoperfusion and a worsening of symptoms. There is some support for this hypothesis when the physical stress is an orthostatic challenge. During postural stress, patients with CFS showed high rates of neurally mediated hypotension (9, 10), increased heart rates (11, 12), and decreased stroke volumes when compared with healthy controls (12) and reported that their symptoms get worse (9). Considering our finding that patients with CFS showed cognitive impairments relative to controls (5), we speculated whether this result might relate to a reduced cardiovascular response. An association between autonomic dysfunction and cognitive impairment has been drawn (13), but little empiric data exist to examine the relation between blood pressure regulation and cognitive function. Therefore, a major purpose of this study was to test the hypothesis that patients with CFS would show a decreased blood pressure response to the mental stress of cognitive challenge.

Psychosomatic Medicine 63:756 –764 (2001)

EXERCISE AND CFS RESPONSE TO COGNITIVE STRESS Because we collected cardiovascular data contiguous with cognitive testing before and after exercise, we were able to evaluate the role of exercise in influencing both cardiovascular and cognitive function. Our specific hypothesis was that the hemodynamic consequences of exercise might uncover subtle autonomic dysfunction and thus magnify an impaired cardiovascular response to mental stress. With exercise as a probe, there was a possibility of finding normal cardiovascular reactivity to stressful challenge in the baseline condition in CFS, but reduced responding when testing was done after exercise. Thus, the second purpose of this study was to determine whether physical exertion would uncover or magnify a diminished cardiovascular response to the mental stress of cognitive challenge. Tracking cardiovascular function during a stressful cognitive challenge—initially, when the cognitive functioning of patients with CFS and control subjects was similar, and then again after exhaustive exercise, when the cognitive function of patients with CFS was reduced—provided data that addressed a third purpose of this study, namely, to determine whether there is a relationship between cardiovascular reactivity and reduced cognitive performance in patients with CFS. METHODS Subject Recruitment The patient group was composed of 19 consecutive white women with CFS who agreed to participate in this study. Patients either were self-referred to the New Jersey CFS Cooperative Research Center or were referred by their physicians. We chose to confine the study to women who were white for two reasons: first, because more than 90% of the patients in our center are women and more than 97% are white; and second, to reduce heterogeneity of results due to gender factors. The women were all studied during the luteal phase of their menstrual cycles—again, to reduce heterogeneity of results due to cyclical effects on cardiovascular regulation (14). All the patients with CFS met the 1988 Center for Disease Control working case definition of CFS (1). Thus, every patient reported a new onset of severe fatigue that had persisted or relapsed for at least 6 months, caused at least a 50% reduction in activity from premorbid levels, and could not have been attributed to known medical conditions. The patient also reported eight or more of the following: low-grade fever, sore throat, painful lymph nodes, muscle weakness, myalgia, symptom worsening after mild exercise, headaches, migratory arthralgia, neuropsychological complaints, sleep disturbance, and sudden onset of symptoms (1). Psychiatric exclusions based on a computerized psychiatric diagnostic interview (15) included schizophrenia, bipolar or eating disorders occurring at any time in the patient’s life; substance/alcohol abuse occurring in the 10 years before illness onset; or any Axis I psychiatric diagnosis occurring in the 5 years before illness onset. Additional exclusions were illness duration longer than 6 years, loss of consciousness for more than 10 minutes, the diagnosis of hypertension, or use of antihypertensives, antibiotics, daily inhalers, steroids, or benzodiazepines. Twenty healthy, sedentary normotensive controls were recruited

Psychosomatic Medicine 63:756 –764 (2001)

by advertisements in area newspapers and flyers placed in public places within an approximately 30-mile radius of the CFS Center. These subjects were matched to patients with CFS for age, gender, race, and education (see Table 1), and no significant differences were seen in these variables for the CFS and control groups (all comparisons p ⬎ .05). “Sedentary” was defined as working in an occupation that required minimal physical labor and not participating in physical exercise for more than one session per week. Controls were excluded if they had a history of medical illness, a major psychiatric diagnosis within the 5 years before the study as determined by the diagnostic psychiatric interview, or were taking any medications other than birth control pills.

Instrumentation The cognitive test battery used in this study has been fully described elsewhere (5). Briefly, it consisted of the following tests: (1) verbal forms of the SCWT (16), which is a measure of cognitive speed and disinhibition (2); the paper and pencil Symbol Digit Modalities Test (SDMT) (17), which is a measure of psychomotor speed and vigilance (3); an oral version of the Trail Making Test (18), which is a measure of attention/concentration; and (4) an oral subtraction test, which required serially subtracting 13 from 100. During the cognitive testing periods, a Finapres (Model 2300, Datex-Ohmeda, Tewkesbury, MA) device was used to continuously measure beat-by-beat arterial blood pressure. The Finapres uses the finger-volume clamp principle described by Penaz (19). Also, a 12-lead ECG was used to continuously monitor heart rate during baseline and the cognitive tests (Q4000, Quinton Instrument Company, Seattle, WA). The analog signals from the ECG and Finapres were digitized at 200 Hz on a personal computer system by use of the Snap-Master data acquisition program (HEM Data Co., Southfield, MI). Pulse wave and ECG signal analysis were performed off-line by use of a program implemented in S-Plus software (Statistical Sciences, Seattle, WA). Self-ratings of tiredness, energy, calmness, and tension were assessed by use of the AD ACL, which is a test for the transitory levels of these arousal states (20). The BDI was used to ascertain level of depressed mood (21). Current CFS symptom severity was estimated via a questionnaire that contained a Likert severity score for each of the symptoms listed in the 1988 case definition for CFS (1). These ratings were summed to quantify the extent of the subjects’ symptoms. The exercise test and metabolic measurements have been described previously (5). The exercise test was a graded walking test to TABLE 1.

Subject Characteristics and Baseline Data for the Groups (Means ⴞ SEM)

Variable Age (yr) Height (cm) Weight (kg) BDI AD ACL Energy Tiredness Tension Calmness

CFS Group (N ⫽ 19)

Control Group (N ⫽ 20)

34 ⫾ 2 164.3 ⫾ 6.0 67.3 ⫾ 14.4 14.7 ⫾ 6.8a

34 ⫾ 2 162.2 ⫾ 1.5 61.8 ⫾ 2.7 1.6 ⫾ 0.7

7.9 ⫾ 0.7a 14.8 ⫾ 1.0a 8.4 ⫾ 0.7 12.0 ⫾ 1.0

13.4 ⫾ 0.7 6.5 ⫾ 0.5 7.2 ⫾ 0.6 12.0 ⫾ 0.9

a Significantly different (p ⬍ .05) when compared with the control group.

757

J. J. LAMANCA et al. exhaustion performed on a motorized treadmill. During this test, the subjects breathed through a two-way breathing valve attached to a mask that covered their noses and mouths (Hans-Rudolph, Kansas City, MO). Expired air was analyzed for ventilatory and metabolic ˙ O2peak, via a Q-PLEX I metabolic system, and variables, including V cardiac activity was recorded by use of the Q4000 ECG monitor (Quinton).

the cognitive test battery was administered by use of the same protocol described previously. After completing this post-treadmill cognitive test battery, the subjects completed the AD ACL. On the second day, subjects arrived at the laboratory at the same time as day 1 and completed the same protocol as described above for the preexercise testing session. Briefly, they completed the CFS Symptoms questionnaire and the AD ACL. A blood sample was taken. They performed the cognitive test while physiological variables were being recorded, and then they completed the AD ACL.

Procedures Subjects were asked to abstain from performing any heavy physical work or exercise and ingesting caffeine for 24 hours, and eating or smoking for 3 hours before presenting for laboratory testing, to eliminate the effect of these variables on heart rate, blood pressure, and fatigue. The subjects reported to the human performance laboratory at the same time on 2 consecutive days. On the first day, after signing an informed consent and completing a short questionnaire that ascertained whether they had complied with the pretesting conditions stated above, the subjects completed the BDI, CFS Symptoms Questionnaire, and the AD ACL. They then sat quietly for 30 minutes, after which a blood sample was taken by venipuncture from an antecubital vein. The results from the blood samples are reported elsewhere (22). Briefly, many of the immunological variables significantly increased (p ⬍ .05) with exercise, but no significant differences (p ⬎ .05) were found between the CFS and control groups for any of the lymphocyte fractions or cytokines measured at any time during the experiment (22). Next the subjects were prepped for ECG and blood pressure measurements. The Finapres cuff was placed on the index finger of the subject’s nondominant hand and was kept at approximately heart level by use of an adjustable table. The height of this table was marked for each subject to standardize the arm height during retesting, and subjects were instructed to refrain from moving this arm or hand during testing. All cognitive testing was performed in a quiet temperature-controlled room in the sitting position. The test administrator was seated across a table from the subject. A laboratory technician and the electronic equipment were behind and out of sight of the subject. The technician controlled the data acquisition and marked events during the testing sessions. The subjects performed the cognitive test battery in the following order: SCWT, SDMT, Trail Making Test, and the subtraction test. Presentation of instructions for the SCWT was followed by 5 minutes of quiet rest while baseline physiological measurements were recorded, and then the test periods were conducted. After the SCWT and before each additional test, the subject was read standardized instructions by the test administrator and performed the test immediately with no rest period. Each test instruction period took approximately 30 seconds, except for those for the SDMT, which took approximately 2 minutes. The total cognitive test battery time including baseline was approximately 16 minutes. Physiological data were recorded continuously throughout baseline and the cognitive testing. The times of the beginning and ending of the performance of each individual test in the cognitive test battery were marked on the digitized data. After the cognitive test battery, the subjects completed the AD ACL. The subjects were then prepared with instruments for collection of heart rate and respiration data during the exercise test. After reaching the peak stage where the subjects indicated they could go no further, the treadmill was stopped. A second blood sample was taken 4 minutes after exercise. The subjects then completed the AD ACL, and, within 10 minutes after completion of the treadmill test,

758

Data Reduction and Statistical Analysis Our initial evaluation of the cardiovascular data used an analysis of variance for repeated measures based on the baseline data and on data acquired during each instruction period and test period during the cognitive test battery. There were no consistent within-group differences in cardiovascular responses across the different test and instruction periods and no significant group interactions over time. Because these findings as well as findings that the cardiovascular variables did not return to baseline levels during the instructional periods, we elected to treat the cognitive test battery as a compound stressor and combined the scores during the cognitive test battery as follows. Baseline cardiovascular data were expressed as an average of the last two minutes of the rest period before a cognitive test battery was started. During the cognitive test battery, the cardiovascular data were averaged across each individual test performance. These mean heart rate and blood pressure values for each individual cognitive test were averaged to give one value for the complete cognitive test battery. The cardiovascular data were analyzed by use of a 3 (cognitive test trials: pre- , immediate- , and 24-hour postexercise test) ⫻ 2 (cardiovascular values during baseline and cognitive test period) ⫻ 2 (groups: CFS and control) mixed ANOVA (general linear models; SAS System for Solaris Release 6.12; SAS Institute; Gary, NC) and the Greenhouse-Geisser probability levels. The primary focus of this study was cardiovascular responses of CFS to cognitive stress. Accordingly, the main analysis compared patients with CFS to healthy controls. However, the literature indicates that depression can have a significant effect on an individual’s stress response (3), and the patients with CFS in the present study had higher BDI values when compared with controls (14.7 ⫾ 1.6 vs. 1.6 ⫾ 0.7, p ⬍ .01). Therefore, we used a subsequent analysis to examine the effects of a depressive mood state on the cardiovascular reactivity to the cognitive challenges. For this analysis, we divided the CFS group into those with high (ⱖ16; N ⫽ 9; mean 20.2 ⫾ 1.3) and low (⬍16; N ⫽ 10; mean 9.7 ⫾ 1.3) BDI scores on the basis of a median split. This stratagem was previously used in a study of CFS (23). We used the change scores on the cardiovascular values during the cognitive tests and a 3 (groups: high-BDI CFS, low-BDI CFS, and control) ⫻ 3 (cognitive test trials: pre- , immediate- , and 24-hour postexercise test) mixed ANOVA and the Greenhouse-Geisser probability levels. Linear regression was used to examine the relationship between cardiovascular changes (cognitive test mean minus baseline mean) and each of the following: cognitive test performance, the BDI scores ˙ O2peak. Also, linear regression was of the patient with CFS, and V used to see whether CFS symptom severity scores and AD ACL bipolar energy scores could predict CFS cardiovascular responses during cognitive testing. Data are reported as mean and standard error of the mean.

Psychosomatic Medicine 63:756 –764 (2001)

EXERCISE AND CFS RESPONSE TO COGNITIVE STRESS RESULTS Cardiovascular Responses During the Cognitive Test Batteries The means for the cardiovascular variables during baseline and cognitive testing are presented in Table 2. The ANOVA with repeated measures revealed a significant effect of performing the cognitive test battery on HR [F(1,37) ⫽ 95.23, p ⬍ .0001], SBP [F(1,37) ⫽ 175.74, p ⬍ .0001], DBP [F(1,37) ⫽ 235.29, p ⬍ .0001], MAP [F(1,37) ⫽ 236.01, p ⬍ .001], and PP [F(1,37) ⫽ 61.67, p ⬍ .0001]. However, significant interactions indicated that the CFS group had a lower response for HR [F(1,37) ⫽ 8.47, p ⬍ .01], SBP [F(1,37) ⫽ 7.74, p ⬍ .01], DBP [F(1,37) ⫽ 7.41, p ⬍ .01], and MAP [F(1,37) ⫽ 11.80, p ⬍ .001] during the cognitive test sessions when compared with the controls. Figure 1 illustrates the group differences in reactivity for HR, SBP, and DBP responses (ie, cognitive test period mean minus baseline mean) across three testing trials. Also, there was a tendency for the subjects with CFS to have a lower PP response during the cognitive testing when compared with the control group, but this did not reach statistical significance [F(1,37) ⫽ 4.06, p ⬎ .05]. Significant cognitive test trial by cardiovascular reactivity interactions were seen for HR [F(2,74) ⫽ 5.16, p ⬍ .01], SBP [F(2,74) ⫽ 5.71, p ⬍ .01], DBP [F(2,74) ⫽ 7.81, p ⬍ .001], MAP [F(2,74) ⫽ 12.00, p ⬍ .0001], and PP [F(2,74) ⫽ 3.34, p ⬍ .05]. Tests of simple main effects indicated that, regardless of group, reduced responses were seen during the immediate postexercise cognitive test trial when compared with the responses seen during the pre- and 24-hour postexercise cogniTABLE 2.

tive test trials. However, the absence of interactions among cognitive test trials, cardiovascular reactivity, and groups for HR [F(2,74) ⫽ .35, p ⬎ .68], SBP [F(2,74) ⫽ 1.05, p ⬎ .35], DBP [F(2,74) ⫽ .53, p ⬎ .58], MAP [F(2,74) ⫽ .54, p ⬎ .59], and PP [F(2,74) ⫽ 1.41, p ⬎ .25] indicated that the CFS and control groups’ cardiovascular reactivity values were not significantly attenuated by different amounts during the immediate postexercise cognitive test session. Postexercise cardiovascular reactivity to stress can be influenced by such variables as exercise intensity and duration and the subject’s fitness level (24). As we have reported elsewhere, the groups’ aerobic fitness ˙ O2peak during the graded levels, as determined by V exercise test, were not different (27.7 ⫾ 1.6 ml/kg per minute for CFS vs. 30.4 ⫾ 1.0 ml/kg per minute for the control; p ⬎ .10) (5). Also, as an indicator of exercise intensity, there was no difference in the percentage of predicted HRpeak [ie, 210 ⫺ (age ⫻ 0.65)] achieved during the exercise test (93 ⫾ 2% for CFS and 97 ⫾ 1% for the control group; p ⬎ .05). However, the total time of exercise was significantly higher for the control group compared with CFS (19.9 ⫾ 1.1 and 15.7 ⫾ 1.2 minutes, respectively; p ⬍ .05). Therefore, we repeated the ANOVA, using exercise time as a covariate, but found no differences in the group and cardiovascular reactivity interactions from those reported above. Cardiovascular Reactivity and Depressive Mood Table 3 contains the mean cardiovascular reactivity values (cognitive test mean ⫺ baseline mean) across the cognitive test trials for the control, high-BDI CFS,

Cardiovascular Variables at Baseline and During Performance of Cognitive Tests in Patients with CFS and Heathy, Sedentary Controls (Mean ⴞ SEM) Pretreadmilla

Variable HR (bpm)b SBP (mm Hg)b DBP (mm Hg)b MAP (mm Hg)b PP (mm Hg)b

Posttreadmill

24 hr Posttreadmill

Time

Baseline CTB Baseline CTB Baseline CTB Baseline CTB Baseline CTB

CFS

Control

CFS

Control

CFS

Control

79 ⫾ 3 84 ⫾ 3c 128.0 ⫾ 3.0 139.2 ⫾ 3.2c 73.0 ⫾ 1.5 81.1 ⫾ 1.9c 91.6 ⫾ 1.7 101.1 ⫾ 2.2c 55.0 ⫾ 2.2 58.2 ⫾ 2.3

75 ⫾ 2 83 ⫾ 3 128.0 ⫾ 2.7 147.5 ⫾ 3.4 71.6 ⫾ 1.5 83.1 ⫾ 1.9 90.6 ⫾ 1.8 105.6 ⫾ 2.4 56.3 ⫾ 2.2 64.4 ⫾ 2.9

98 ⫾ 3 101 ⫾ 3c,d 120.7 ⫾ 3.7 129.5 ⫾ 3.9c,d 74.1 ⫾ 2.0 79.3 ⫾ 2.5c,d 90.4 ⫾ 2.5 96.3 ⫾ 2.9c,d 46.5 ⫾ 2.3 50.2 ⫾ 2.4d

102 ⫾ 2 107 ⫾ 3d 121.8 ⫾ 3.5 135.9 ⫾ 3.4d 71.0 ⫾ 2.1 79.6 ⫾ 2.0d 87.7 ⫾ 2.3 98.0 ⫾ 2.3d 50.9 ⫾ 2.4 56.3 ⫾ 2.2d

79 ⫾ 2 83 ⫾ 2c 116.3 ⫾ 3.3 130.1 ⫾ 4.0c 67.0 ⫾ 1.7 74.6 ⫾ 2.2c 84.2 ⫾ 2.2 92.9 ⫾ 2.7c 49.4 ⫾ 2.1 55.5 ⫾ 2.4

78 ⫾ 2 84 ⫾ 3 121.7 ⫾ 2.4 139.8 ⫾ 2.4 69.2 ⫾ 1.7 78.9 ⫾ 1.6 87.2 ⫾ 1.6 99.7 ⫾ 1.7 52.5 ⫾ 1.9 60.9 ⫾ 2.1

a Pretreadmill, posttreadmill, and 24 hr posttreadmill refer to the cognitive tests given immediately before, immediately after, and 24 hours after an incremental treadmill exercise test to exhaustion. b Statistically significant (p ⬍ .05) time effect of performing the cognitive tests. c Statistically significant (p ⬍ .05) lower change from baseline when compared with control. d Statistically significant (p ⬍ .05) reduced responses when compared with the responses seen during cognitive testing pre- and 24-hour posttreadmill exercise.

Psychosomatic Medicine 63:756 –764 (2001)

759

J. J. LAMANCA et al. ⬍ .05] reactivity was seen in the low-BDI CFS group when compared with the control group. For the high-BDI CFS group, a trend toward lower SBP [F(1,36) ⫽ 4.07, p ⬍ .06] and DBP [F(1,36) ⫽ 3.66, p ⬍ .07] reactivity was seen when compared with the control group. There was a significant time effect for HR [F(2,72) ⫽ 4.32, p ⬍ .03], SBP [F(2,72) ⫽ 5.31, p ⬍ .01], DBP [F(2,72) ⫽ 7.94, p ⬍ .001], MAP [F(2,72) ⫽ 10.41, p ⬍ .01], and PP [F(2,72) ⫽ 3.27, p ⬍ .05] with all of the reactivity values being attenuated immediately after exercise and returning back to pre-exercise levels 24 hours later. However, the lack of group-by-time interactions HR [F(4,72) ⫽ 1.13, p ⬎ .34], SBP [F(4,72) ⫽ 1.71, p ⬎. 15], DBP [F(4,72) ⫽ 1.99, p ⬎. 14], MAP [F(4,72) ⫽ 1.33, p ⬎ .27], and PP [F(4,72) ⫽ 1.32, p ⬎. 27] indicated that the groups’ cardiovascular reactivity did not change differently across the three cognitive trials. No significant relationship was seen between the BDI scores of the patients with CFS and any of the cardiovascular responses during the cognitive test trials (p ⬎ .05 for all correlations).

Cardiovascular Reactivity and Cognitive Performance Fig. 1.

HR, SBP, and DBP reactivity for the CFS and healthy, sedentary control groups when performing cognitive tests before, immediately after, and 24 hours after incremental treadmill exercise to exhaustion (mean ⫾ SEM). Reactivity scores were calculated by subtracting the cardiovascular value during baseline from the value obtained during the cognitive test period. a ⫽ Significantly different when compared with control (p ⬍ .05); b ⫽ significantly different from pre-exercise values (p ⬍ .05).

and low-BDI CFS groups. When the three groups were compared, significant main effects were seen in the reactivity values for HR [F(2,36) ⫽ 4.18, p ⬍ .03], SBP [F(2,36) ⫽ 3.82, p ⬍ .04], DBP [F(2,36) ⫽ 3.78, p ⬍ .04], and MAP [F(2,36) ⫽ 5.84, p ⬍ .01] but not for PP [F(2,36) ⫽ 1.99, p ⬎ .15]. Tests of simple main effects revealed no significant differences in HR [F(1,36) ⫽ 0.10, p ⬎ .05], SBP [F(1,36) ⫽ 0.09, p ⬎ .05], DBP [F(1,36) ⫽ 0.12, p ⬎ .05], and MAP [F(1,36) ⫽ 0.15, p ⬎ .05] reactivity to the cognitive testing when the highand low-BDI CFS groups were compared. Lower reactivity was seen for HR and MAP in both the high- (HR [F(1,36) ⫽ 6.18, p ⬍ .05]; MAP [F(1,36) ⫽ 6.13, p ⬍ .05]) and low- (HR [F(1,36) ⫽ 4.84, p ⬍ .05]; MAP [F(1,36) ⫽ 9.19, p ⬍ .05] BDI CFS groups when compared with the control group. Also, lower SBP [F(1,36) ⫽ 5.95, p ⬍ .05] and DBP [F(1,36) ⫽ 6.01, p

760

Complete results of the cognitive testing and questionnaire data have been presented elsewhere (5). The following is a brief summary of results. No group differences were seen on the cognitive test battery preexercise, but immediately after and 24 hours after exercise, the CFS cognitive test scores for the SDMT and two parts of the Stroop (ie, Stroop Word Test and Stroop Color Test) were significantly less than those for controls (p ⬍ .05). To evaluate the relationship between cognitive performance and cardiovascular reactivity, we calculated standard scores for the Stroop and SDMT, which are the segments of the cognitive test battery where group performance differences were seen after the exercise (5). Next, we created a cognitive test performance score for each of the three testing sessions by calculating means from these standard scores. Linear regression revealed a positive relationship between the SBP change score (ie, cognitive test period SBP mean minus baseline SBP mean) and the standard test performance score during the immediate (r ⫽ .51, p ⬍ .02) and 24-hour (r ⫽ .48, p ⬍ .04) postexercise cognitive test sessions. Also, a positive correlation was seen between the MAP change values and cognitive performance during all three cognitive testing sessions (rs ⫽ .56, .50, and .58, all relationships p ⬍ .03).

Psychosomatic Medicine 63:756 –764 (2001)

EXERCISE AND CFS RESPONSE TO COGNITIVE STRESS TABLE 3.

Average Group Cardiovascular Reactivitya to Cognitive Stress Pre-, Immediately Post-, and 24 Hours Post–Incremental Treadmill Exercise to Exhaustion (mean ⴞ SEM) Reactivity Values Groups

Low BDI CFS (N ⫽ 10) High BDI CFSc (N ⫽ 9) Controld (N ⫽ 20) b

HR (bpm)

SBP (mm Hg)

DBP (mm Hg)

MAP (mm Hg)

PP (mm Hg)

4.1 ⫾ 0.6 3.5 ⫾ 1.1e 7.1 ⫾ 0.9

10.8 ⫾ 2.1 11.7 ⫾ 2.3 17.2 ⫾ 1.5

6.6 ⫾ 1.1 7.3 ⫾ 1.5 9.9 ⫾ 0.6

7.68 ⫾ 1.6 8.4 ⫾ 2.1e 12.6 ⫾ 1.1

4.2 ⫾ 1.1 4.5 ⫾ 1.2 7.3 ⫾ 1.2

e

e

e

e

a

Cardiovascular reactivity values were calculated by subtracting the baseline mean value from the cognitive test mean value. Low BDI CFS ⫽ CFS patients with BDI scores ⬍ 16. c High BDI CFS ⫽ CFS patients with BDI scores ⱖ 16. d Control ⫽ heathy, sedentary control subjects. e Statistically significant (p ⬍ .05) when compared with control. b

Cardiovascular Reactivity and Illness Severity Because preexercise cardiovascular responses of patients with CFS were not different from their 24-hour postexercise responses, we used the averaged preexercise and 24-hour postexercise scores when computing correlations between symptom severity and SBP response to cognitive stress. The results of this analysis are displayed in Figure 2. Lower SBP reactivity to cognitive testing was associated with higher ratings of CFS symptom severity (r ⫽ ⫺ .62, p ⬍ .01). The baseline AD ACL values are presented in Table 1, and the results of the individual subscales of the AD

ACL (ie, energy, tiredness, calmness, and tension) have been previously reported (5). The following is a brief summary. At no time during the experiment did the groups indicate significant differences in their ratings of calmness and tension. Throughout the experiment, the CFS group consistently indicated less energy and more tiredness when compared with the control group, both before and after exercise and cognitive testing. To gain a more complete indication of energetic arousal than is available using the energy subscale of the AD ACL (20), we computed a bipolar energy score using the AD ACL subscale data. This bipolar energy score is derived by combining the energy ratings with inverted tiredness ratings. Therefore, in the present study, we calculated the CFS subjects’ bipolar energy scores on the first AD ACL (17.6 ⫾ 1.4), which was completed before cognitive testing and exercise. A significant positive correlation was seen between the AD ACL bipolar energy scores and SBP response during the preexercise cognitive test session (r ⫽ .46, p ⬍ .05). ˙ O2peak No significant relationship was seen between V and any of the cardiovascular responses during the cognitive test trials (p ⬎ .05 for all correlations). DISCUSSION

Fig. 2.

The relationship between averaged before and 24 hours after incremental treadmill exercise scores on a CFS symptom severity questionnaire and SBP reactivity during performance of cognitive tests. The symptom severity questionnaire consisted of a Likert severity score for each of the symptoms listed on the 1988 Center for Disease Control case definition of CFS.

Psychosomatic Medicine 63:756 –764 (2001)

The main finding of this study was a diminished cardiovascular response to stressful cognitive testing in patients with CFS when compared with healthy, sedentary controls. This diminished stress response was seen consistently in CFS across three separate testing sessions. Also, significant negative correlations between self-ratings of CFS symptom severity and cardiovascular responses were seen. These data agree with studies that have shown differences between the physiological systems involved in the stress response of healthy controls and subjects with CFS (7, 11, 25, 26). However, our data seem to be in opposition to those studies that have indicated sym-

761

J. J. LAMANCA et al. pathetic hyperactivity in CFS (27), as well as with those that have shown no abnormalities in CFS autonomic and cardiovascular stress reactivity (28). One reason for these differences among studies may relate to the type of stressor used. For instance, De Becker et al. (27) concluded, primarily on the basis of results from head-up tilt testing, that sympathetic hyperactivity exists in patients with CFS. However, head-up tilt produces a significant shift of blood volume from the thoracic region to the legs and involves a vastly different response from that elicited with cognitive stressors like those used in the present study. During passive tilt, there is a need for sympathetic activation, parasympathetic withdrawal, and baroreflex activity to maintain normal cerebral blood flow. Autonomic blockade studies indicate that cognitive effort elicits primarily sympathetic activation (29) and does not involve large blood volume shifts. Because there was a positive relationship between the level of cardiovascular reactivity and performance in the CFS group, one may speculate that a decreased level of effort could account for low cardiovascular responses. However, in our earlier report, we noted that patients with CFS only had problems on some of the tests in the cognitive test battery and were normal in others. This suggests that their effort level was appropriate. In addition, we have reported that the CFS and control groups’ cognitive performance was not different during the preexercise cognitive tests (5), and, at that time, the cardiovascular responses of the CFS group were lower than those of the control group. Thus, at least at this point in time, the reduced cardiovascular responses to stress does not seem to be due to lack of effort. Abnormal autonomic and cardiovascular responses can be precipitated by inactivity and cardiovascular deconditioning (30). Freeman and Komaroff (11) found a relationship between inactivity and autonomic dysfunction in CFS. In the present study, we controlled for the inherent inactivity in CFS by comparing our patients with sedentary controls. Also, we found no differences in the fitness levels of our groups as measured by peak oxygen uptake and no relationship between this variable and cardiovascular reactivity. Depression has also been shown to affect the stress response (3). However, our analysis could not provide any evidence that the results of this study were a function of underlying depressed mood. If depression was a significant factor in producing our results, we would expect to have found an orderly relation with the greatest effect in the patients with the highest BDI scores. However, regression analysis revealed no significant relationship between cardiovascular reactivity and the BDI scores of the patients with CFS, and we

762

found the opposite result from that expected, with the low-BDI CFS group having the lowest blood pressure responses to cognitive stress when compared with the control group. We hypothesized that exercise would exacerbate any differences in CFS and control cardiovascular stress response, but this did not happen. Consistent with the literature, we saw a general attenuation in the cardiovascular responses to cognitive stress immediately after exercise (31). However, without a no-exercise control group, we cannot be sure how much of this attenuated response is due to exercise and how much is due to physiological habituation caused by repeated testing over a short period of time. Regardless, a lack of significant interactions indicated that CFS responses were not suppressed to a greater degree than those seen in controls. One might be tempted to speculate that there is a threshold of cardiovascular response needed for optimal cognitive performance, and the decreased postexercise cognitive scores seen in the CFS group may have been a result of their cardiovascular response being below this threshold. Most likely this was not the case, because 24 hours after exercise, the CFS group’s cognitive performance was still decreased when compared with controls, but their cardiovascular responses were equivalent to preexercise values, when group cognitive performance was not different. Another hypothesis one might make from these data are that the low blood pressure responses of the subjects with CFS may have resulted in low blood perfusion of the brain, which, in turn, could have resulted in the decreased cognitive performance seen. However, because cardiovascular hyporesponsiveness was seen during the pre-exercise cognitive test battery, when the CFS and control cognitive scores were not different, drawing a causal relationship between hyporesponsiveness and cognitive performance would be highly speculative. One study has investigated heart rate responses in CFS during the performance of the Baddeley grammatical reasoning task (32). The cognitive tests were given with and without background noise. These researchers found that the cognitive performance of subjects with CFS decreased with the addition of noise stress. They saw higher heart rates in their patients with CFS both at baseline and during testing when compared with healthy controls. However, the heart rates of patients with CFS did not increase more than those of the controls during task performance. Unfortunately, the researchers did not measure blood pressure (32). The similarity in this study and ours relates to the decrement in cognitive performance. In the work of Beh et al. (32), the cognitive performance of subjects with CFS did not differ from that of controls until an addi-

Psychosomatic Medicine 63:756 –764 (2001)

EXERCISE AND CFS RESPONSE TO COGNITIVE STRESS tional stressor (ie, noise) was added. In our study, the cognitive performance of both patients with CFS and controls were the same before exercise, and differences were not manifest until after the additional stress of exercise. We found a negative relationship between CFS symptom severity on the experimental days and SBP increases during the cognitive stressors. This indicates that the “sicker” the individual is feeling, the less her stress reactivity. This conclusion is supported by data on individual energy levels among patients with CFS, as determined from the AD ACL. When these were low at the time of testing, the SBP response to stress was also low. Data from our laboratory on a group of Persian Gulf War veterans with chronic fatigue support these results, because that patient group also exhibited a decreased blood pressure reactivity to cognitive stress tests that was associated with increased ratings of fatigue (33). One potential limitation of our work is that, as might be expected, patients often will not participate in research when they feel very sick. Therefore, had we been able to evaluate patients at the time of extremely severe symptoms, the abnormal cardiovascular stress responses might have been worse. Recently, it has been reported that differences in the immune function were seen in patients with CFS outside but not during laboratory conditions (34). Also, frequent periods of abnormal electrocardiograms have been seen from ambulatory Holter monitoring in patients with CFS (35). In summary, this study found an abnormal cardiovascular stress response in women with CFS and that exercise did not magnify this effect. That the poorest responders had the worst symptoms suggests that the individual patient’s response to stress plays a role in the common complaint of symptom worsening after stress. Thus, it is possible that stress provocation and reactivity testing could be used as an independent measure of efficacy in therapeutic trial research in this disabling condition. This work was supported by National Institutes of Health Center Grant U01 AI-32247. We thank Theresa L. Policastro, Nancy Hill, and Tracy A. Sakowski for help in the preparation of this manuscript and subject recruitment. REFERENCES 1. Holmes GP, Kaplan JE, Gantz NM, Komaroff AL, Schonberger LB, Straus SE, Jones JF, Dubois RE, Cunningham-Rundles C, Pahwa S. Chronic fatigue syndrome: a working case definition. Ann Intern Med 1988;108:387– 89. 2. Cleare AJ, Wessely SC. Chronic fatigue syndrome: a stress disorder? Br J Hosp Med 1996;55:571– 4.

Psychosomatic Medicine 63:756 –764 (2001)

3. Chrousos GP, Gold PW. The concepts of stress and stress systems disorders: overview of physical and behavioral homeostasis. JAMA 1992;267:1244 –52. 4. Sternberg E. Hypoimmune fatigue syndromes: diseases of the stress response? J Rheumatol 1993;418 –21. 5. LaManca JJ, Sisto SA, DeLuca J, Johnson SK, Lange G, Pareja J, Cook S, Natelson BH. Influence of exhaustive treadmill exercise on cognitive functioning in chronic fatigue syndrome. Am J Med 1998;105:59S– 65S. 6. Sisto SA, Tapp WN, LaManca JJ, Ling W, Korn LR, Nelson AJ, Natelson BH. Physical activity before and after exercise in women with chronic fatigue syndrome. Q J Med 1998;91: 465–73. 7. Cordero DL, Sisto SA, Tapp WN, LaManca JJ, Pareja JG, Natelson BH. Decreased vagal power during treadmill walking in patients with chronic fatigue syndrome. Clin Auton Res 1996;6:329 –33. 8. Blackwood S, MacHale S, Power M, Goodwin G, Lawrie S. Effects of exercise on cognitive and motor function in chronic fatigue syndrome and depression. J Neurol Neurosurg Psychiatry 1998;65:541– 6. 9. Bou-Holaigah I, Rowe P, Kan J, Calkins H. The relationship between neurally mediated hypotension and the chronic fatigue syndrome. JAMA 1995;274:961–7. 10. Schondorf R, Benoit J, Wein T, Phaneuf D. Orthostatic intolerance in the chronic fatigue syndrome. J Auton Nerv Syst 1999; 75:192–201. 11. Freeman R, Komaroff A. Does the chronic fatigue syndrome involve the autonomic nervous system? Am J Med 1997;102: 357– 64. 12. LaManca JJ, Peckerman A, Walker J, Kesil W, Cook S, Taylor A, Natelson BH. Cardiovascular response during head-up tilt in chronic fatigue syndrome. J Clin Physiol 1999;19:111–20. 13. Karas B, Grubb BP, Boehm K, Kip K. The postural orthostatic tachycardia syndrome: a potentially treatable cause of chronic fatigue, exercise intolerance, and cognitive impairment in adolescents. Pacing Clin Electrophysiol 2000;23(3):344 –51. 14. Girdler SS, Light KC. Hemodynamic stress responses in men and women examined as a function of female menstrual cycle phase. Int J Psychophysiol 1994;17(3):233– 48. 15. Marcus S, Robbins LW, Bucholz K. Quick diagnostic interview schedule III-R version 1. St. Louis: University School of Medicine; 1990. 16. Golden CJ. Stroop color and word test: a manual for clinical and experimental uses. Wood Dale, IL: Stoelting Company; 1978. 17. Smith A. Symbol digit modalities test manual. Los Angeles, CA: Western Psychological Services; 1982. 18. Ricker JH, Axelford BN, Houlter BD. Clinical validation of the oral trail making test. Neuropsychiatry Neuropsychol Behav Neurol 1996;9:50 –3. 19. Penaz J. Photoelectric measurement of blood pressure, volume and flow in finger. Digest Int Conf Med Biol Engineer Dresden 1973;104. 20. Thayer RE. The biopsychology of mood and arousal. New York: Oxford University Press; 1989. 21. Beck AT, Ward CH, Mendelson M. An inventory for measuring depression. Arch Gen Psychiatry 1961;4:561–71. 22. LaManca JJ, Sisto SA, Zhou XD, Ottenweller JE, Cook S, Peckerman A, Zhang Q, Denny TN, Gause WC, Natelson BH. Immunological response in chronic fatigue syndrome following a graded exercise test to exhaustion. J Clin Immunol 1999;19: 135– 42. 23. Marshall P, Forstot M, Callies A, Peterson P, Schenck C. Cognitive slowing and working memory difficulties in chronic fatigue syndrome. Psychosom Med 1997;59:58 – 66.

763

J. J. LAMANCA et al. 24. Roy M, Steptoe A. The inhibition of cardiovascular responses to mental stress following aerobic exercise. Psychophysiology 1991;28:689 –700. 25. Scott LV, Medbak S, Dinan TG. Blunted adrenocorticotropin and cortisol responses to corticotropin-releasing hormone stimulation in chronic fatigue syndrome. Acta Psychiatr Scand 1998; 97:450 –7. 26. Sisto S, Tapp WN, Drastal S, Bergen M, DeMasi I, Cordero DL, Natelson BH. Vagal tone is reduced during paced breathing in patients with chronic fatigue syndrome. Clin Auton Res 1995; 5:139 – 43. 27. De Becker P, Dendale P, De Meirleir K, Campine I, Vandenborne K, Hagers Y. Autonomic testing in patients with chronic fatigue syndrome. Am J Med 1998;105:22S– 6S. 28. Duprez D, De Buyzere M, Drieghe B, Vanhaverbeke F, Taes Y, Michielsen W, Clement DL. Long- and short-term blood pressure and RR-interval variability and psychosomatic distress in chronic fatigue syndrome. Clin Sci 1998;94:57– 63. 29. Sherwood A, Allen M, Obrist A, Langer A. Evaluation of betaadrenergic influences on cardiovascular and metabolic adjustments to physical and psychological stress. Psychophysiology 1986;23:89 –104. 30. Lamb L, Johnson R, Stevens P, Welch B. Cardiovascular decon-

764

31.

32. 33.

34.

35.

ditioning from space cabin simulator confinement. Aerospace Med 1964;420 – 8. Boone J, Probst M, Rogers M, Berger R. Post exercise hypotension reduces cardiovascular responses to stress. J Hypertens 1993;11:449 –53. Beh H, Connelly N, Charles M. Effect of noise stress on chronic fatigue syndrome patients. J Nerv Ment Dis 1997;185:55– 8. Peckerman A, LaManca J, Smith S, Taylor A, Tiersky L, Pollet C, Korn L, Hurwitz B, Ottenweller J, Natelson B. Cardiovascular stress responses and their relation to symptoms in Gulf War Veterans with fatiguing illness. Psychosom Med 2000;62: 509 –16. Gupta S, Aggarwal S, Starr A. Increased production of interleukin-6 by adherent and non-adherent mononuclear cells during ⬍natural fatigue⬎ but not following ⬍experimental fatigue⬎ in patients with chronic fatigue syndrome. Int J Mol Med 1999;3: 209 –13. Lerner A, Goldstein J, Chang C, Zervos M, Fitzgerald J, Dworkin HJ, Lawrie-Hoppen C, Korokin SM, Brodsky M, O’Neill W. Cardiac involvement in patients with chronic fatigue syndrome as documented with Holter and biopsy data in Birmingham, Michigan, 1991–1993. Infect Dis Clin Pract 1997;6:327–33.

Psychosomatic Medicine 63:756 –764 (2001)