Left Ventricular, Peripheral Vascular, and Neurohumoral Responses to Mental Stress in Normal Middle-Aged Men and Women
Reference Group for the Psychophysiological Investigations of Myocardial Ischemia (PIMI) Study
Background The normal cardiovascular response to mental stress in middle-aged and older people has not been well characterized.
Methods and Results We studied 29 individuals 45 to 73 years old (15 women, 14 men) who had no coronary risk factors, no history of coronary artery disease, and a negative exercise test. Left ventricular (LV) volumes and global and regional function were assessed by radionuclide ventriculography at rest and during two 5-minute standardized mental stress tasks (simulated public speaking and the Stroop Color-Word Test), administered in random order. A substantial sympathetic response occurred with both mental stress tests, characterized by increases in blood pressure, heart rate, rate-pressure product, cardiac index, and stroke work index and rises in plasma levels of epinephrine and norepinephrine but not β-endorphin or cortisol. Despite this sympathetic response, LV volume increased and ejection fraction (EF) decreased secondary to an increase in afterload. The change in EF during mental stress varied among individuals but was associated positively with changes in LV contractility and negatively with baseline EF and changes in afterload. EF decreased >5% during mental stress in 12 individuals and >8% in 5; 3 developed regional wall motion abnormalities.
Conclusions Mental stress in the laboratory results in a substantial sympathetic response in normal middle-aged and older men and women, but EF commonly falls because of a concomitant rise in afterload. These results provide essential age- and sex-matched reference data for studies of mental stress–induced ischemia in patients with coronary artery disease.
Although the cardiovascular changes that accompany physical exercise have been well characterized, considerably less is known about the changes that occur during mental stress. Many studies have documented that mental stress commonly provokes symptomatic or silent myocardial ischemia in patients with CAD.1 2 3 4
Mental stress induced in the laboratory3 5 can produce a significant sympathetic response that differs in certain respects from the sympathetic response of physical stress. With mental stress, there are relatively greater increases in plasma epinephrine than norepinephrine,6 7 8 and there is a more rapid rise in blood pressure and a lesser increase in heart rate7 compared with physical stress. Noninvasive studies of LV function by radionuclide techniques or echocardiography have generally found an increase or no significant change in EF during mental stress in most normal individuals but a decrease in EF and an appearance of regional wall motion abnormalities in as many as 70% of patients with CAD.1 2 4 9 10 11 12 13 However, only limited information is available about mental stress–induced changes in LV function in normal middle-aged or older individuals who are similar in age to patients with CAD. The definition of what constitutes a “normal” LV response to mental stress has generally been extrapolated from studies of physical exercise, even though changes in LV loading conditions and contractility differ markedly.
The Psychophysiological Investigations of Myocardial Ischemia (PIMI), a multicenter study in 196 patients with known CAD, has sought to identify the psychological and physiological factors that influence the expression of cardiac ischemia during physical and mental stress.14 To establish valid criteria for mental stress–induced ischemia in the PIMI population, a reference group of normal age-matched men and women who had no coronary risk factors and no clinical evidence of CAD was studied. Under a protocol identical to the PIMI study, the reference group underwent RVG during two mental stress tasks—simulated public speaking and the Stroop Color-Word Test—to accurately quantify the “normal” changes that occur in LV function during mental stress.
Inclusion and Exclusion Criteria
Thirty-two individuals were recruited for this study, 8 from each of the 4 PIMI Study sites. The protocol was approved by the Investigational Review Board at each site, and all participants provided informed consent. Participants met all of the following inclusion criteria: (1) no history of CAD, cardiac arrest, or chest pain syndrome; (2) no history of cardiac medications; (3) absence of coronary risk factors (no history of hypertension, diabetes, peripheral vascular disease, or hypercholesterolemia [total cholesterol ≤250 mg/dL and LDL cholesterol ≤160 mg/dL] and of first-degree relatives with myocardial infarction or coronary death before age 50 years); (4) normal blood pressure (systolic <140 mm Hg and diastolic <90 mm Hg); (5) currently a nonsmoker; and (6) a normal resting ECG and a normal exercise treadmill test at ≥85% of predicted maximum heart rate within the past year. Exclusion criteria included (1) the presence of neurological disease (eg, dementia, Parkinson's disease, multiple sclerosis, history of stroke with residual deficit, peripheral neuropathy, or chronic pain disorder), (2) taking of certain medications that could not be discontinued and that could influence the hemodynamic response to stress or pain perception (psychotropic agents or antidepressants, sympatholytics, central nervous system–acting α-agonists, anti-inflammatory agents, analgesics, theophylline, β-adrenergic agonists, or β-adrenergic blockers, including eye drops for glaucoma), and (3) inability to read or communicate in English.
Studies were performed in the morning in all participants. The Stroop test was demonstrated, and the participant was asked to practice. An intravenous line was inserted, an AECG monitor (Applied Cardiac Systems) was attached, and a blood pressure cuff (Dinamap) was applied for automatic recording. Blood was drawn for in vitro labeling of red blood cells with [99mTc]pertechnetate (Ultratag, Mallinckrodt Medical). After a 30-minute rest period, blood samples were drawn for resting epinephrine, norepinephrine, β-endorphin, and cortisol levels, three measurements of blood pressure and heart rate were obtained, and the 99mTc-labeled red blood cells were reinjected (20 to 30 mCi).
The participants were then transferred to a semiupright seated position, and a gamma scintillation camera was positioned approximately in the 40° left anterior oblique view to provide the best separation between left and right ventricles. A baseline RVG (2-minute acquisition) and a 12-lead ECG were obtained. Two mental stress tests—the Stroop Color-Word Test and a simulated public speaking task—were then performed in random order, each lasting 5 minutes. During each task, two 2-minute RVG images were acquired, beginning 30 seconds and 3 minutes after the start of the task. At 1 minute into each task, blood was drawn for epinephrine and norepinephrine levels, and at the completion of the task, the levels of these neurohormones were measured again, along with β-endorphin and cortisol levels. β-Endorphin and cortisol levels were again measured 10 minutes after the second task. Blood pressure, heart rate, and a 12-lead ECG were recorded at 30 seconds after the start of each mental stress task and every minute thereafter. After the mental stress tasks were completed, a 5-mL blood sample was drawn and placed in a Petri dish, and the radioactivity was measured with the same camera/collimator system as used for the RVG acquisition. In addition, a radioactive marker was placed on the chest over the LV, the camera was moved to the anterior position, and a static image was acquired of the LV and the marker to allow calculation of the attenuation distance from the center of the LV to the chest wall. These data were used to calculate LV volumes by a previously validated count-based method.15
Public Speaking Task
Each participant was asked to give a 5-minute speech about a real-life situation involving a close friend or relative receiving poor care by the staff in a hospital or nursing home. The participant was given 1 minute to prepare the speech and was instructed to talk as if actually involved in the situation and to express his/her feelings.
Stroop Color-Word Test
Each participant performed a computerized version of the Stroop Color-Word Test using a color video display and a three-key mouse. One of three words appeared in the middle of the screen—“red,” “green,” or “blue”—and each of these words was printed along the bottom of the screen. However, the color of the print did not match the word spelled. The participant was asked to match the color of the word in the middle of the screen with the spelling of the color at the bottom by pressing the appropriate mouse key. The display changed quickly, forcing the subject to make rapid selections in the face of confusing information. This computerized version of the Stroop test provided standardized administration and scoring (including presentation of instructions) and automatic titration of difficulty based on level of performance.
Analysis of RVGs
Radionuclide images were transferred from each clinical site to the Radionuclide Core Laboratory at Johns Hopkins by floppy disk and read directly or translated through a Sudbury Image Center (Sudbury Systems) into a Sophy NXT nuclear medical computer (Sopha Medical Systems Inc) for analysis and subsequent archiving on optical disk.
Global EF was measured with a commercially available fully automatic method (Sopha Medical). To verify the correctness of the automated method and to provide measurements of EF when the automatic program was unable to identify the LV or track the LV edges, a manual analysis of EF was performed on the resting RVG of every participant. For this analysis, the end-diastolic image was identified, and a region of interest was constructed around the LV. This region of interest was applied to all the images in the RVG study to identify the end-systolic image, and a separate LV region was then constructed for end systole, along with a background region inferior and lateral to the LV. The automatic EF was accepted as correct if the manual EF was within 5 EF units. If it was not, the manual EF was repeated, and if this determination was within 5 EF units, the automatic EF was accepted. If it was not, but was within 5 EF units of the second manual result, the second manual result was reported. If all three measurements were >5 EF units from each other, a third manual EF was done and reported as the “correct” EF. For this study, the fully automatic EF was accepted in 93% of prestress and 82% of mental stress RVGs.
Regional wall motion was scored subjectively for each RVG on a four-point scale (3, dyskinetic; 2, akinetic; 1, hypokinetic; and 0, normal) for each of four predefined segments (septal, inferoapical, low lateral, and high lateral). For this analysis, the baseline RVG, the two RVGs acquired during public speaking, and the two RVGs acquired during the Stroop test were shown together on the video screen in random order. Studies were mixed together with similar studies from the PIMI population (all of whom had proven CAD). The evaluator was therefore blinded to whether the subject was normal or had CAD and to whether an individual RVG was obtained at baseline or during mental stress. A new or worse wall motion abnormality was considered to occur during a mental stress task only if the abnormality was present in both of the 2-minute images acquired during that task.
Calculation of LV volumes was performed by a validated count-based method using individualized attenuation correction, described in detail elsewhere.15 The LV EDV was obtained from the ratio of the attenuation-corrected end-diastolic count rate from the RVG to the count rate per milliliter from a venous blood sample obtained in each subject and counted at the clinical site. ESV was calculated from the measured EF and EDV. Stroke volume was calculated as the difference between the EDV and ESV, and cardiac output was the product of stroke volume and heart rate measured during the RVG. Measurements of cardiac volumes by this method, as well as other related count-based methods, have been validated under resting15 16 and exercise17 18 conditions.
Rate-pressure product was calculated as SBP×heart rate. Cardiac volumes were expressed in absolute terms or as indexes (per square meter of body surface area). Mean arterial pressure was calculated as [SBP+2×DBP]/3. Total SVR was calculated as mean arterial pressure×80/cardiac output; SWI as stroke volume index×SBP; and LV contractility index as SBP/ESV index.19 Effective Ea, a measure of vascular load that accounts for pulsatile flow, was calculated as end-systolic pressure divided by stroke volume, approximated by [2×systolic pressure+diastolic pressure]/3×stroke volume.20
Analysis of AECGs
The 4-hour AECG recordings were analyzed with a CardioData Mk4 playback system with modified software21 in the AECG Core Laboratory. Both the technician and the physician who reviewed the tapes were unaware of whether the subject was normal or had CAD (as a participant in the PIMI Study) or whether the Speech or Stroop task was performed first. An ischemic episode was defined as transient ST-segment deviation ≥1.0 mm lasting ≥1.0 minute.
Measurement of Biochemical Variables
Blood samples were immediately placed on ice; the plasma was separated within 30 minutes, frozen at −85°C, and sent to the PIMI Biochemistry Core Laboratory, where analyses were run without knowledge of patient identity or the sequence of blood collection. All samples from a given subject were analyzed in the same batch in duplicate. Norepinephrine and epinephrine concentrations were measured by reverse-phase, ion-pair high-performance liquid chromatography in combination with computer-controlled cation-enrichment precolumn and a three-electrode electrochemical detector. The mean interassay coefficient of variation for pooled samples was 8.6% for norepinephrine (mean concentration of 770 pg/mL) and 12.6% for epinephrine (mean concentration of 109 pg/mL). β-Endorphin concentrations were assayed by radioimmunoassay after affinity-gel extraction. Mean interassay coefficient of variation for pooled samples was 13.3% (mean concentration of 23.2 pmol/L). Sensitivity (maximum bound +3 SD) was about 3 pg/L. Cortisol concentrations were assayed by a coated-tube radioimmunoassay. The mean interassay coefficient of variation was 6.1%.
Multivariable linear regression was used to assess the association of changes in response to mental stress with age and sex. Each model included a change variable with age (continuous) and sex (categorical). Separate analyses were performed for the Speech and Stroop tests. Correlation coefficients were calculated for all change variables with the changes in EF and plasma catecholamine levels. Both univariable and multivariable linear regression were performed to assess the association of the derived hemodynamic variables with change in EF. Separate regression models were fitted for Speech, Stroop, and maximum EF change during either mental stress task.
Study Group Characteristics
Of the 32 subjects recruited and studied, 3 were excluded because of an elevated resting blood pressure and/or evidence of myocardial disease. The exclusions were a 70-year-old man with a baseline EF of 46% and a resting regional wall motion abnormality, a 76-year-old man with a baseline SBP of 165 mm Hg and a resting regional wall motion abnormality, and a 71-year-old woman with a baseline SBP of 176 mm Hg.
The characteristics of the study group are presented in Table 1⇓. A treadmill test was done in each individual and was negative for ischemia.
Heart Rate and Blood Pressure Changes During Mental Stress
Increases in blood pressure and heart rate occurred during both mental stress tasks, somewhat more so during simulated public speaking (Speech) than during the Stroop Color-Word Test (Table 2⇓). Rate-pressure product increased 45.8% during Speech and 30.7% during Stroop (P<.05) (Fig 1⇓). Significant age-related differences were observed in resting heart rate and rate-pressure product (lower in subjects 45 to 54 years old than in subjects 55 to 64 or 65 to 73 years old) (Table 2⇓). Age-related differences also were observed in the amount of increase in SBP and rate-pressure product during Speech (less in younger subjects). Sex differences were present in resting heart rate (lower in men) and the increase in DBP during Speech (greater in men).
AECG Change During Mental Stress
ST-segment depression occurred in only 1 subject, a 62-year-old woman, for 5 minutes during Speech only (maximal ST depression, 1.3 mm). The EF increased 3% and regional wall motion remained normal in this individual.
LV Function During Mental Stress
Mental stress caused significant changes in LV function (Table 3⇓). Mean EDV increased slightly during both Speech and Stroop (Fig 1⇑). ESV increased (P<.001) and EF decreased (P<.001) during Stroop. Stroke volume did not change significantly during either mental stressor, but cardiac index increased, mediated by an increase in heart rate. The increase in EDV during Speech was significantly age-related (greatest in 45- to 54-year-old subjects). Although resting EDV and stroke volume were higher in men than women, the changes in LV function during either mental stressor were not sex-related.
Derived Hemodynamic Variables During Mental Stress
LV SWI increased substantially during both Speech and Stroop (P<.001, Table 4⇓, Fig 1⇑). Overall, mean SVR did not change significantly, although individual responses were variable, with some subjects showing an increase and others a decrease. Similarly, mean LV contractility index did not change significantly, although some individuals had an increase and others a decrease. During Speech, contractility index increased in 22 subjects (12 men, 10 women) and decreased in 6 (2 men, 4 women). During Stroop, contractility index increased in 11 subjects and decreased in 16. On average, there was a modest increase in contractility index during Speech and a small decrease during Stroop (P<.05 for Speech versus Stroop). Effective Ea, reflecting vascular loading conditions, increased significantly (P<.001) during both Speech and Stroop. Women had a higher mean Ea than men at rest (P=.03), probably because of a smaller stroke volume. There were no significant differences in SWI, SVR, or contractility index related to age or sex.
Neurohumoral Changes During Mental Stress
There were significant increases in plasma epinephrine and norepinephrine levels during both Speech and Stroop (Table 5⇓). Epinephrine levels peaked early during mental stress and were higher at 1 minute than at 5 minutes during both Speech and Stroop. The increase in epinephrine was somewhat greater during Speech than Stroop (P=NS). Men had an approximately twofold greater rise in mean plasma epinephrine than women during both Speech and Stroop (P<.03) (Fig 2⇓). Norepinephrine levels, in contrast, tended to be higher at 5 minutes than at 1 minute during mental stress. The increase in norepinephrine level was somewhat greater during Speech than Stroop and was greater in men than women (P<.05 for Speech). A ≥50% increase in norepinephine level occurred in 64% of men versus only 27% of women during Speech (P=.04). Overall, there were no significant age-related differences in either epinephrine or norepinephrine during mental stress, but older women (55 to 73 years old, n=8) had higher maximum norepinephine levels than younger women (45 to 54 years old, n=7) during both Speech (518±147 versus 347±126 pg/mL, P=.03) and Stroop (503±147 versus 339±100 pg/mL, P=.03), whereas baseline levels did not differ. Epinephrine levels during mental stress were not significantly different in older and younger women.
The peak serum norepinephrine level was significantly correlated during Speech with the maximum SBP (r=.50, P=.006), rate-pressure product (r=.44, P=.02), SWI (r=.49, P=.008), stroke volume (r=.35, P=.07), and cardiac output (r=.59, P=.004) but not with heart rate, SVR, LV contractility index, Ea, EDV, ESV, or change in EF. During Stroop, peak norepinephrine level was correlated with SBP (r=.39, P=.04), rate-pressure product (r=.33, P=.08), SWI (r=.39, P=.05), LV contractility index (r=.35, P=.08), and cardiac output (r=.41, P=.03). Peak epinephrine level was not significantly correlated with any of those variables during either Speech or Stroop. The change in norepinephrine level was correlated with the changes in SBP (r=.47, P=.01) and rate-pressure product (r=.43, P=.02) during Speech and with the changes in SBP (r=.37, P=.05) and SVR (r=.35, P=.07) during Stroop. The change in epinephrine level was correlated only with the change in SWI (r=.38, P=.05) during Speech and the change in LV contractility index (r=.36, P=.07) during Stroop.
Serum β-endorphin level at baseline averaged 5.5±3.1 pmol/L. During Speech, the mean plasma level did not change significantly (+0.15±24 pmol/L), but during Stroop, it fell 0.9±2.2 pmol/L (P=.04). The decrease during Stroop was seen in women (−1.7±2.3 pmol/L) but not in men (+0.04±1.7 pmol/L) (P=.03 for sex difference). At 10 minutes after mental stress, mean β-endorphin level had fallen significantly to 3.7±1.9 pmol/L (P=.007), and this decrease occurred equally in men and women. There were no age-associated differences in β-endorphin levels.
Plasma cortisol level averaged 12.4±3.8 mg/dL at rest, 13.0±4.0 mg/dL during Speech, and 13.4±4.3 mg/dL 10 minutes after the final mental stress test. These increases from resting level were not significant with either mental stress task, and no age- or sex-related differences were observed.
Changes in LV EF and Regional Wall Motion
Although mean EF decreased during both Speech (−2.4±6.5%, P<.1) and Stroop (−4.4±4.5%, P<.001), the individual responses during mental stress were variable and ranged from a 15% decrease to a 21% increase (Fig 3⇓). The variability appeared to be greater in women than men.
A fall in EF of >5% is commonly used in the literature to identify those patients with CAD who develop myocardial ischemia during mental stress.1 2 12 Use of this criterion would have resulted in 8 (28%) of our subjects being called abnormal during Speech and 10 (34%) during Stroop (12 [41%] on either task, and 6 [21%] on both, Table 6⇓). Table 6⇓ shows the number of individuals identified as “abnormal” based on a decrease in EF of >5%, >6%, >7%, or >8%. As the criteria become more stringent, fewer normal people are identified as abnormal, but even at a threshold of >8%, 1 of 14 men (7%) and 4 of 15 women (27%) had an abnormal EF response to one or both mental stressors.
A regional wall motion abnormality was identified at baseline in 1 subject (a 50-year-old woman, EF of 73%), but the abnormality did not worsen during mental stress. New wall motion abnormalities during mental stress were seen in 3 subjects, 1 during Speech and 2 during Stroop (Fig 3⇑). The 1 man with a new wall motion abnormality also had a decrease in EF of 12% during Stroop; in the 1 woman with a new wall motion abnormality during Stroop, EF decreased 4%, and in the 1 woman with a new wall motion abnormality during Speech, EF decreased 15%.
Follow-up 99mTc-sestamibi perfusion studies were done in all 6 individuals in whom EF decreased >8% and/or a regional wall motion abnormality developed during mental stress (1 man, 5 women). Perfusion imaging (tomographic in 5, planar in 1) was combined with exercise testing in 5 people and dipyridamole infusion in 1 person (a woman with a recently fractured hip). The scintigraphic images were read as normal in all cases.
Determinants of EF Change During Mental Stress
A decrease in EF during mental stress should be attributable to an increase in afterload placed on the LV, a decrease in contractility, or both. Overall, Ea, a measure of afterload, increased during mental stress as EF decreased (Fig 1⇑). Mean changes in SVR and contractility index were not statistically significant. The increase in Ea was somewhat greater during Speech than during Stroop. Despite this, EF decreased less during Speech. This finding may have been related to a somewhat greater increase in contractility during Speech, which tended to offset the increase in afterload.
To analyze this further, the changes in EF in individual subjects were related to the changes that occurred in afterload and contractility indexes as well as to baseline levels of EF. By multiple linear regression analysis, the maximal change in EF during either mental stress task was inversely related to the pre–mental stress EF value (r=−.59, P=.0008) but was not related to age; sex; the levels of SBP or DBP, heart rate, or rate-pressure product at baseline or during mental stress; or the changes in these variables during mental stress. The inverse relation between change in EF and pre–mental stress EF was significant for all subjects and for women alone (r=−.79, P=.0005) but not for men alone (r=−.19, P=.52), possibly related to the narrow range of pre–mental stress EF values.
The maximal change in EF during mental stress was related inversely to the concomitant changes in both SVR (r=−.56, P=.0018) and Ea (r=−.59, P=.0009) and directly to the change in LV contractility index (r=.72, P=.0001) (Fig 4⇓). In a multivariate analysis, the change in EF was significantly associated with pre–mental stress EF (P=.05) and with changes in LV contractility index (P=.0001), SVR (P=.05), and Ea (P=.001). Although the changes in SVR and Ea were highly correlated with one another (r=.75), all four variables were significant in the model and together explained 84% of the variance in the change in EF during mental stress. Changes in norepinephrine and epinephrine levels were not significant in this analysis.
Our results demonstrate that mental stress created in the laboratory in normal middle-aged and older men and women by simulated public speaking or the Stroop Color-Word Test may cause a substantial sympathetic response, characterized by increases in serum epinephrine and norepinephrine levels (but not in cortisol or β-endorphin levels) and rises in blood pressure, heart rate, rate-pressure product, cardiac index, and SWI. Despite this sympathetic response, LV EDV and ESV rise and EF falls, secondary to an increase in LV afterload, which is reflected in a rise in mean effective Ea. The change in EF among individuals is quite variable but can be explained in large measure by the level of the baseline EF and by the changes in SVR, mean Ea, and LV contractility index that occur during mental stress. The change in EF in normal individuals thus appears to represent a balance between the effects of increased afterload and increased contractility on the LV.
Unique Features of Our Study
This study represents one of the few published studies characterizing the integrated response of the normal cardiovascular system to mental stress. Although other studies have used noninvasive imaging techniques to examine the EF response to mental stress, ours is among the first to put the EF response in the context of neurohumoral and peripheral vascular changes that occur at the same time. Another unique feature of our study is that we have focused our attention on strictly defined normal people >45 years of age and have studied both men and women. Previous studies have used primarily young men as normal control subjects, whereas it is the normal response in middle-aged and older people that must be understood in order to provide valid reference information for studies in patients with CAD. Although our study population was not large, it exceeded the size of all other normal reference groups reported in the literature.1 2 4 9 10 11 12 13 Our results offer the insight that in middle-aged and older people, mental stress causes a decrease in EF, on average, due to an increase in afterload that is not offset by a sufficient increase in contractility.
Since our participants did not undergo coronary arteriography, we cannot entirely rule out the possibility that some had occult CAD and that ischemia contributed to the overall decline in EF during mental stress. However, our subjects were carefully chosen for this study. No participant had a history of CAD, none had significant coronary risk factors, and all had a normal exercise test. The few individuals who developed either a new wall motion abnormality during mental stress (n=1), a fall in EF of >8% (n=3), or both (n=2) underwent subsequent noninvasive evaluation for occult CAD, and in all cases the results were negative. We therefore believe that if occult CAD was present, it did not contribute significantly to our overall results.
Differences Between Speech and Stroop
Our results demonstrated that Speech produced a greater sympathetic response than the Stroop test in this normal middle-aged and older population. The amount of “stress” produced by a given mental stress task is dependent on environmental factors, including the subject's state of baseline sympathetic arousal, the precise way in which the stressor is applied, and the subject's degree of involvement and background of prior experiences relevant to the particular stressor. To control for these variables, we used the same detailed standardized protocols at each of the four clinical sites.14 In addition, the study personnel from each site trained together, and site visits were conducted during practice runs before the actual study was begun. Nevertheless, it is unknown to what extent our results are applicable to other populations (eg, younger individuals or racial minorities) or whether modifications of our mental stress protocols would have altered the results.
At baseline, male participants had a significantly lower heart rate, larger EDV and stroke volume, and a lower Ea than female participants. The differences in heart volumes probably relate to the larger male body size, whereas the lower heart rate in men is compatible with a lower baseline sympathetic tone. During mental stress, however, male participants had an approximately 2- to 2.5-fold greater increase in epinephrine and norepinephrine levels than female participants, indicating a greater adrenergic response to the same stimulus. These results are consistent with the findings of previous studies.22 23 In women, the adrenergic response to mental stress increases at menopause,24 and in men, estrogen has been shown to reduce the adrenergic response to stress.25
Comparison With Previous Studies
Previous studies have reported that mental stress results in increases in heart rate, blood pressure, and cardiac output and a decrease in peripheral resistance.2 3 4 10 11 12 26 27 28 In addition, activation of the sympathetic nervous system has been found, with a marked rise in plasma catecholamine levels.6 11 29 30
Kaji et al11 used serial M-mode echocardiography to characterize changes in LV pump function and peripheral hemodynamics in 20 young, healthy men during mental arithmetic. They found that heart rate, blood pressure, cardiac output, stroke volume, EF, and LV contractility index increased. Plasma epinephrine increased, but norepinephrine did not. Decreases were found in peripheral resistance and LV ESV, whereas LV EDV did not change. The differences between our study and that of Kaji et al11 (ie, we found overall a decrease in EF, increase in EDV and ESV, increase in Ea, and no change in LV contractility index) are most likely due to differences in the study populations. Our data suggest that women may have a less intense β-adrenergic response to mental stress than men, and inclusion of women in our study may have reduced the overall positive inotropic changes. In addition, our data also suggest that older individuals may have a greater increase in LV afterload (Ea) during mental stress than younger individuals. Since half of our study population was >54 years old, our results may reflect a greater influence of afterload over inotropy for the population as a whole.
What Constitutes an “Abnormal” Change in EF During Mental Stress?
As the criterion for an abnormal EF response during mental stress is made more stringent (ie, a greater decrease in EF is required to define “abnormal”), the number of people from normal reference populations identified as abnormal becomes smaller (ie, the test becomes more specific). Several previous studies have defined a fall in EF during mental stress of >5% as abnormal.1 2 12 Rozanski et al1 reported that 0 of 12 normal subjects had a decrease in EF of >5%, but Bairey et al2 found that in 4 of 18 normal subjects, the EF decreased >5%. The latter study also reported that 3 of 18 had a decrease in EF of >6%, 3 of 18 a decrease of >7%, and 2 of 18 a decrease of >8%. Ironson et al4 found that 1 of 9 normal subjects had a decrease of ≥7%. In each of these studies, the control populations consisted of relatively young individuals, almost entirely male. In our own study, 12 individuals had a decrease in EF of >5% on either Speech or Stroop tasks, 8 a decrease of >6%, 7 a decrease of >7%, and 5 a decrease of >8% (Table 6⇑). Interestingly, large decreases in EF occurred more frequently in the female participants: 4 of 15 women had a decrease of >8%, compared with 1 of 14 men. Thus, if a fall in EF of >8% is used as the criterion for abnormality, a specificity of 93% is found for men, whereas the specificity for women is only 73%.
The reason for the large (>8%) decrease in EF during mental stress in 1 man and several women is not entirely clear. The change in EF in our study participants was correlated in multivariate analysis with resting EF level and with changes in LV contractility and afterload (SVR and Ea). All of the individuals with a large decrease in EF had a resting EF of >70%, all had a pronounced decrease in contractility index (range, 0.53 to 5.45 mm Hg·mL−1·m−2), and all had large increases in SVR (range, 89 to 330 dynes·s·cm−5) and Ea (range, 0.21 to 0.58 mm Hg/mL). A pattern of high resting EF combined with a decrease in contractility and an increase in afterload in these subjects during mental stress is consistent with high resting sympathetic tone and withdrawal of β-adrenergic tone and/or an increase in α-adrenergic tone during mental stress. A high resting sympathetic state could represent an individual's characteristic baseline or a temporary arousal related to the laboratory setting. In either event, the level of resting sympathetic tone may have an important impact on the subsequent physiological response to stress.
Knowledge of the resting EF and the changes that occurred in LV contractility index, SVR, and Ea allowed prediction of the change in EF during mental stress with a high degree of confidence. A multivariate regression model explained 84% of the variability in EF change during mental stress. However, it should be recognized that the derived variables of LV contractility index, SVR, and Ea each contain measures of LV volume that are themselves correlated with EF (ie, the change in EF was correlated with the changes in cardiac output [r=.46], stroke volume [r=.65], and ESV [r=.78]) and even use the same volumes in the calculations (eg, ESV is used in the calculation of EF, SVR, LV contractility index, and Ea). The association between change in EF and changes in these derived variables, although they make physiological sense, could nonetheless be spurious in whole or in part. Each derived variable has been validated as a useful index of physiological change in previous studies, but we had no radionuclide-independent measure of change in contractility or afterload to relate to EF change in this study.
Regional Wall Motion Abnormalities
New regional wall motion abnormalities were identified during mental stress in 3 study subjects. In each case, exercise radionuclide perfusion studies were subsequently done to look for occult CAD, but all were normal. In previous studies, regional wall motion abnormalities were reported in 1 of 12 subjects by Rozanski et al1 and in 0 of 9 subjects in Ironson et al.4 Since our population was older, we may have inadvertently included individuals with occult CAD despite negative treadmill tests and exercise perfusion studies. Alternatively, the new wall motion abnormalities may have represented a false-positive finding related to inaccuracies in image acquisition or interpretation. In our study, wall motion was read subjectively, and the interpreter was blinded to whether an image corresponded to baseline or mental stress and whether it belonged to a normal individual or to a patient with CAD.
If a decrease in EF of >8% is combined with development of a new wall motion abnormality as a criterion for an abnormal mental stress radionuclide study, 1 of 14 men and 5 of 15 women in our normal population would be classified as abnormal (specificities of 93% and 67%, respectively, P=.08 for men versus women). Thus, although the RVG appears to provide adequate specificity in men for studying mental stress–induced myocardial ischemia, the results in women may be more problematic and suggest that a high false-positive rate could interfere with the accurate identification of mental stress–induced ischemia in women with CAD. Further investigation of sex differences appears to be warranted.
Selected Abbreviations and Acronyms
|CAD||=||coronary artery disease|
|DBP||=||diastolic blood pressure|
|LV||=||left ventricular, left ventricle|
|SBP||=||systolic blood pressure|
|SVR||=||systemic vascular resistance|
|SWI||=||stroke work index|
This study was funded by the National Heart, Lung, and Blood Institute, Behavioral Medicine Research Group, Division of Epidemiology and Clinical Applications, National Institutes of Health, Bethesda, Md, by research contracts HV-18114, HV-18119, HV-18120, HV-18121, and HV-28127. Support of ECG data collection was provided in part by Applied Cardiac Systems, Laguna Hills, Calif; Marquette Electronics, Milwaukee, Wis; Mortara Instrument, Milwaukee; and Quinton Instruments, Seattle, Wash. Dinamap equipment was provided by Critikon Inc, a Johnson & Johnson Company. Michael Eddy (University of Pittsburgh) and Richard Lutz (University of North Carolina) provided Stroop test software, and Dr William Maixner (University of North Carolina) provided software and design of the Marstock sensory threshold test. Some centers had partial support from General Clinical Research Center grants.
Reprint requests to PIMI Clinical Coordinating Center, Maryland Medical Research Institute, 600 Wyndhurst Ave, Baltimore, MD 21210.
*A list of participating centers and investigators appears in Reference 14.
- Received January 2, 1996.
- Revision received June 25, 1996.
- Accepted July 8, 1996.
- Copyright © 1996 by American Heart Association
Krantz DS, Helmers KF, Bairey CN, Nebel LE, Hedges SM, Rozanski A. Cardiovascular reactivity and mental stress-induced myocardial ischemia in patients with coronary artery disease. Psychosom Med. 1991;53:1-12.
Manuck S, Krantz D. Psychophysiologic reactivity in coronary heart disease and essential hypertension. In: Matthew K, Weiss S, Detre TM, Dembroski T, Falkner E, Manuck S, Williams R, eds. Handbook of Stress, Reactivity and Cardiovascular Disease. New York, NY: John Wiley & Sons Inc; 1986.
Legault SE, Freeman MR, Langer A, Armstrong PW. Pathophysiology and time course of silent myocardial ischemia during mental stress: clinical, anatomical, and physiological correlates. Br Heart J. 1995;73:242-249.
Gottdiener JS, Krantz DS, Howell RH, Hecht GM, Klein J, Falconer JJ, Rozanski A. Induction of silent myocardial ischemia with mental stress testing: relation to the triggers of ischemia during daily life activities and the ischemic functional severity. J Am Coll Cardiol. 1994;24:1645-1651.
Goldberg AD, Becker LC, Bonsall R, Cohen JD, Ketterer MW, Kaufmann PG, Krantz DS, Light KC, McMahon RP, Noreuil T, Pepine CJ, Raczynski J, Stone PH, Strother D, Taylor H, Sheps DS, for the PIMI Investigators. Ischemic, hemodynamic, and neurohormonal responses to mental and exercise stress: experience from the Psychophysiological Investigations of Myocardial Ischemia Study (PIMI). Circulation. In press.
Links JM, Becker LC, Shindledecker JG, Guzman P, Burow RD, Nickoloff EL, Alderson PO, Wagner HN Jr. Measurement of absolute left ventricular volume from gated blood pool studies. Circulation. 1982;65:82-91.
Sorensen SG, Ritchie JL, Caldwell JH, Hamilton GW, Kennedy JW. Serial exercise radionuclide angiography: validation of count-derived changes in cardiac output and quantitation of maximal exercise ventricular volume change after nitroglycerin and propranolol in normal man. Circulation. 1980;61:600-609.
Wyns W, Melin J, Dehouck Y, Vanbutsele R, Steels M, Piret L, Detry JM. Radionuclide absolute left ventricular volumes during exercise: validation in normals by Fick measurements. Am J Cardiol. 1982;49:1031. Abstract.
Kelly RP, Ting C-T, Yang T-M, Liu C-P, Maughan WL, Chang M-S, Kass DA. Effective arterial elastance as index of arterial vascular load in humans. Circulation. 1992;86:513-521.
Frankenhaeuser M, von Wright MR, Collins A, von Wright J, Sedvall G, Swahn CG. Sex differences in psychoneuroendocrine reactions to examination stress. Psychosom Med. 1978;40:334-343.
LeBlanc J, Cote J, Jobin M, Labrie A. Plasma catecholamines and cardiovascular response to cold and mental activity. J Appl Physiol.. 1979;47:1207-1211.