| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Department of Clinical Psychology (P.S.), University of Liverpool, Liverpool; Departments of Pharmacology (S.C.S.) and Psychology (S.Z.), University College London, London; Harefield Hospital (G.M.), Middlesex; and the Royal Brompton National Heart and Lung Hospital (J.P.), London, United Kingdom.
Address reprint requests to: Professor Peter Salmon, Department of Clinical Psychology, University of Liverpool, Whelan Building, Brownlow Hill, Liverpool L69 3GB, UK. Email: psalmon{at}liv.ac.uk
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
METHODS: Fifty-six transplantation patients, 66 bypass patients, and 40 patients with heart failure underwent a 10-minute, computer-generated, Stroop color-word conflict test. Heart rate and systolic and diastolic blood pressures were recorded continuously for 1 minute before, during, and 12 minutes after the stressor. Emotional state was measured periodically by questionnaires.
RESULTS: All hemodynamic variables were increased by the Stroop test. There was a pattern of blunted response to the Stroop test after cardiac transplantation, particularly in comparison with bypass patients, and slower recovery in comparison with both control groups. Emotional stress responses were similar in each group.
CONCLUSIONS: This pattern cannot be attributed to the experience of major heart surgery or to cardiac disease. Nor can it be explained by differences in central processing of stress. Correspondingly the changed hemodynamic response to the Stroop test after cardiac transplantation evidently does not affect patients’ emotional responses. The hemodynamic findings are consistent with an increased reliance on hormonal rather than neuronal hemodynamic regulation after cardiac transplantation.
Key Words: heart • transplantation • stress • hemodynamic • emotion.
Abbreviations: CABG = coronary artery bypass graft; ECG = electrocardiogram; HF = heart failure; OCT = orthotopic heart transplant; VDU = video display unit.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
This suggestion is borne out by studies of heart transplant recipients’ response to physical exercise to the extent that when compared with normal, healthy subjects, their maximum heart rate is lower and this chronotropic response is slower in onset and recovery (1, 2). But cardiac output is less affected by transplantation. This is partly because the "Frank-Starling" adjustment ensures that left ventricular stroke volume normally increases in line with ventricular filling (3) and also because hormonal catecholamines increase stroke volume in the absence of any tonic sympathetic activity.
Against this background is extensive evidence that different types of stressful challenge induce different hemodynamic responses (reviewed by Herd, Ref. 4). Of particular importance are transplant patients’ responses to psychological stressors, not least because these occur frequently in daily life and, unlike physical exercise, are usually unpredictable and unavoidable. Therefore, understanding responses to psychological stressors could be important for postoperative rehabilitation and appropriate clinical management.
An early report indicated that neuronal decentralization of the heart of rhesus monkeys virtually eliminated the short-latency increases in heart rate and left ventricular pressure induced by conditioned appetitive or aversive stimuli (5). However, more recent studies of cardiac transplant recipients suggest that the deficits are more modest. Sloan et al. (6) found clear chronotropic, systolic and diastolic blood pressure responses to a mental arithmetic task in a small sample of cardiac transplant recipients. It was also confirmed in another sample that hemodynamic reactivity was less than in either normal subjects or renal transplant patients. Sehested et al. (7) compared responses to an intelligence test between cardiac transplant recipients and age-matched control subjects. The increase in heart rate in the control subjects, which is thought to be neurogenic in origin (8), was absent in the transplant patients. However, changes in systolic and diastolic pressure did not differ between the two groups. This could be because under most conditions, systolic pressure is not dependent on any autonomic innervation of the heart (8) and in OCT patients, diastolic pressure is mainly determined by the (intact) sympathetic innervation of the resistance vessels. Similar findings were reported by Shapiro et al. (9) in a prospective study of cardiac transplant recipients and age- and gender-matched control subjects.
This pattern of evidence suggests that cardiac transplantation impairs the hemodynamic response to a psychological stressor, especially the changes in heart rate. However, comparison with healthy control subjects does not control for long-standing cardiac disease or its enduring consequences, such as physical deconditioning. Nor does it control for the experience of major cardiac surgery. The small sample sizes (5–20 subjects) and the variability of the experimental stressors further weaken conclusions from studies reported until now. Moreover, published research has neglected the possibility that subjects’ perception of the stressor could be altered by cardiac transplantation: After experiencing such major surgery, patients might simply be less disturbed by minor challenges. Alternatively, because emotional reactions (particularly negative emotions) are influenced by interoceptive stimuli (10), an altered hemodynamic stress response could have a secondary impact on the emotional response.
The present study aimed to test whether the hemodynamic response to a psychological stressor is disrupted after cardiac transplantation, specifically that the response is of longer latency and duration than in patients with intact extrinsic innervation. A parallel objective was to test whether patients’ emotional response to the stressor was affected by transplantation and whether this could be related to their hemodynamic status. Heart rate and blood pressure were monitored continuously to ascertain whether there were differences in the extent and/or time course of the response in these different groups of patients. Emotional reactivity to the stressor was also evaluated psychometrically in the same patients. These measurements were carried out in patients who had undergone OCT and were compared with reactivity of two other groups. Patients with end-stage HF, who were awaiting cardiac transplantation, controlled for a history of severe cardiac disease; patients who had undergone CABG surgery controlled for the experience of major cardiac surgery. The stressor was the Stroop color-word conflict procedure, which reliably elicits heart rate, blood pressure, and anxiety responses in healthy young subjects (11–14).
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
General criteria for exclusion were: age less than 20 or more than 75 years; evidence of previous or concurrent psychiatric illness; insufficient understanding of English to complete questionnaires; insulin-dependent diabetes mellitus; and persistent, clinically defined neurological disorder. Criteria for excluding transplant subjects were more than one transplanted organ and coronary artery disease. Patients in the HF group were excluded if there was evidence of renal or hepatic dysfunction. Only two subjects in the OCT group had an infection at the time of testing, one with Epstein-Barr virus and the second with an unidentified virus. Recent infections were also reported in two subjects in the HF group (influenza and a chest infection within 2 days and 2 weeks of the study, respectively).
Subjects were recruited either during their routine attendance at their respective outpatient clinics or by telephone. Of those approached, 15 refused outright to take part (HF, 9; CABG, 4; and OCT, 2), mainly because they felt too unwell; 45 (HF, 18; CABG, 14; and OCT, 13) accepted their appointments but did not attend, usually because of transportation problems; 3 (CABG, 1; HF, 2) were excluded after testing because of technical problems that resulted in a large proportion of missing data. Details of patients for whom data were analyzed are summarized in Table 1. Patients completed questionnaires while awaiting testing and after implantation of a radial arterial catheter into the nondominant arm.
|
After subjects had given informed consent, they were asked about their drug history. They then received an explanation of the procedure before a radial arterial catheter was implanted into the nondominant arm for blood sampling (data reported in Refs. 15 and 18) and recording of blood pressure and heart rate (see below). ECG electrodes were also attached for continuous monitoring (see below). Subjects then remained seated for a minimum of 40 minutes, during which they completed psychological questionnaires (data not reported here). Immediately before the start of the Stroop test, subjects were given a set of headphones (Racal Ampligard AMP4130) to occlude extraneous noise and to sound a buzzer during the test (see below). During all phases of the test the experimenter sat behind a screen so that there was no interaction with the subject except for withdrawal of blood samples. All phases of the Stroop test were fully automated and controlled by microcomputer.
The Stroop Test.
Following on-screen instructions, which subjects could read at their leisure, subjects completed a questionnaire (see below). Subjects then entered an initial relaxation phase (10 minutes), during which a kaleidoscope color sequence was presented on the VDU. This began when a subject pressed one of the buttons on the control panel to signal that they had read the instructions and completed the questionnaire. At the end of this period, subjects repeated the questionnaire and a further set of instructions about the task was provided; again, this was terminated when the subject pressed a control button. The Stroop color-word conflict task then began. Color names were presented on the VDU in either the named or a different color. The task was to press one of two keys on the electronic panel to indicate whether the color name and the color in which the name was displayed were the same or different. The pace of the presentation was 1 word per second so that an entirely correct performance was beyond every subject’s capability. Subjects were told that every mistake would be penalized by addition of an extra word to the test. In fact, the test was always of the same duration (10 minutes). Subjects’ self-esteem was threatened further by telling them that their scores would be compared with those of other subjects at the end of the experiment. The test was followed by a further set of instructions and a third questionnaire. After 2 minutes, recovery continued for another 10 minutes, during which the subject watched the kaleidoscope sequence again. This ended with a final questionnaire.
Hemodynamic Data Recording and Processing.
Recording and initial processing.
The arterial line led to a pressure transducer (Abbott Critical Care Systems: model OJ237-01, lot 80-482SN), from which the signal passed to a Datascope 870 cardiac monitor, to which the ECG electrodes were also connected. This produced two analog outputs, one indicating pressure (at 1 V/100 mm Hg) and the other tracking the ECG signal (at 1 V/mV). These signals were fed into a microcomputer via a 12-bit analog-to-digital converter (configured to map a -5V to +5V range to 12 bits). The pressure signal was sampled at 100 Hz and the ECG signal at 200 Hz. Both the ECG and pressure signal were displayed continuously on the computer screen to allow the experimenter to monitor the patient. The digitized pressure signal was stored on computer disk. Further analysis of this signal began by passing it through a 4253H smoother (using the Minitab program) to remove electronically derived noise (16, 17). A computer algorithm identified the peaks and the troughs in these data and recorded the time and magnitude of each. Artifacts caused mainly by loss of the pressure signal during blood sampling were discarded by identifying pulses that were atypical in shape of the peak-to- peak pressure profile. Transitions between phases of the experiment were recorded.
Mean minute-by-minute scores.
Means of the values from the pressure signal were then determined to produce scores for 1-minute periods (1) at the end of relaxation but before completion of the questionnaires, (2) successively during the 10-minute period of the Stroop test, and (3) for 12 successive minutes after the Stroop test. Mean heart rate for a 1-minute bin was calculated from the peak-to-peak intervals of every pulse that started in that bin. Mean systolic and diastolic pressures were determined across all the peak and trough values within a bin.
Smoothing of hemodynamic variables.
Each variable was smoothed in the following way. Treating the variable as a function of time, f(t), over a fixed period T, it was reexpressed as the sum of harmonic components of the form: ancos(2
nt/T) + bnsin(2nt/T), where an and bn were Fourier coefficients that were computed from f(t) and n was set at 50. Values of each variable for specific points of the response were derived after smoothing. For heart rate and systolic and diastolic blood pressure, values were extracted for the baseline relaxation period (minimum value reached), stress period (maximum value, maximum gradient during the ascent to this value, minimum value reached after the maximum), and recovery period (minimum value reached).
Questionnaires.
Emotional state was measured by the Profile of Mood States (19), which provided separate scores for: tension-anxiety, mental fatigue, anger-hostility, depression-dejection, confusion, and vigor. A total mood disturbance score was calculated by summing all the scales except for vigor, which was subtracted. In addition, self-esteem was assessed by a semantic differential scale including the following items, opposite poles of which were separated by a seven-point scale: hopeless-hopeful, good-bad, unsuccessful-successful, powerful-powerless, important-unimportant, stupid-clever, skillful-unskillful, strong-weak, unconfident-confident, tired-energetic, competent-incompetent, dull-clearheaded, withdrawn-sociable, and contented-discontented. Items were scored in the direction of negative feelings, and scores were summed. The internal consistency of this scale was assessed on each occasion by Cronbach’s 
.
Statistical Analysis
Preliminary analyses assessed the correlation of each hemodynamic measure with age and sex. Only one measure, vigor, differed between the sexes, and some hemodynamic variables were related to age. Relevant variables were therefore adjusted for age or sex by analysis of covariance or, before nonparametric analyses, by linear regression. These adjustments did not materially change any effect; therefore, only adjusted results are presented.
For mean hemodynamic variables that were determined minute by minute, repeated-measures analysis of variance was used to compare groups across successive 1-minute bins. Each analysis confirmed a highly significant change in scores over time (results not reported). Therefore, to examine differences in each phase of the experiment, separate analyses of the baseline (one-way), Stroop period (two-way repeated measures over minutes 1–10), and recovery period (two-way repeated measures over minutes 11–22) were performed. Linear trends were fitted to successive minutes in the analyses of the Stroop test and recovery periods, and significance was assessed by comparison with the linear trend-by-subject within-group term. Age was a covariate where appropriate. Stroop values were analyzed both with and without adjustment by analysis of covariance for baseline values; recovery values were analyzed both with and without adjustment for covariance with the final minute of the Stroop test. To check whether drug use was associated with different hemodynamic responses, we also compared patients receiving or not receiving each drug in any group in which a minimum of 20% and maximum of 80% of patients received that drug. Therefore, for each main type of drug listed in Table 1 that met this criterion (ie, all except immunosuppressants), analyses of variance and covariance were performed, similar to those described above, and included a between-subject factor to identify patients receiving the drug.
Variables derived from smoothed hemodynamic data were analyzed by Kruskal-Wallis nonparametric analysis of variance. Those that were correlated with age were first adjusted for age by linear regression. Maximum values obtained during the Stroop test were also examined after adjustment by linear regression for covariance with minimum baseline values, and minimum Stroop values were also examined after adjustment for the maximum Stroop value. For questionnaires, two-way repeated-measures analyses of variance compared changes over the four measurement times and between groups.
To calculate p values, degrees of freedom were first adjusted to allow for any missing data. Then degrees of freedom that referred to repeated measures in successive bins were adjusted by using the Greenhouse-Geisser 
when the variance-covariance matrix departed significantly from sphericity. Significant effects were explored by post hoc comparisons, using the appropriate error term from the analysis of variance or covariance or, for smoothed hemodynamic variables, Mann-Whitney U tests.
The criterion for significance was p < .05. Analyses were conducted by using SPSS 9.0 and Genstat 5.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Psychological Responses
For the semantic differential (self-esteem) scale, Cronbach’s reliability) exceeded 0.88 at each time of measurement. Semantic differential scores increased in response to the Stroop test ( Figure 1; F(3,474) = 26.22, p < .001). Total mood disturbance also increased (F(3,477) = 55.17, p < .001), which reflected increases in all negative moods (minimum F(3,477) = 14.81, p < .001). Both the semantic differential and total mood disturbance changed similarly in each group (group by time: F(6,474) = 0.95, 0.81, p > .05). Only vigor showed an interaction (F(6,477) = 3.51, p < .01): Whereas HF was unchanged, levels in the OCT and CABG groups declined over testing without a Stroop response. For total mood disturbance, the main effect of group (F(2,159) = 4.45, p < .01) confirmed that mood was worse throughout testing in HF patients than in either the OCT (t = 2.98, p < .01) or CABG (t = 2.35, p < .05) groups. This effect was attributable to differences in fatigue (F(2,159) = 6.42, p < .01).
|
|
|
|
|
Smoothed Data.
Groups differed in the maximum values of diastolic pressure after controlling for baseline (Table 3). This gradient was lesser in the OCT group than in the CABG group (U = 1205, p = .001) and lesser in the HF than the CABG group (U = 960, p < .05). Gradients for the HF and OCT groups were similar. For heart rate (Table 3), the ascent to the maximum reached a lesser gradient in the OCT group than in both the HF (U = 678, p = .001) and CABG (U = 832, p < .001) groups.
After adjustment for the maximum values reached, the subsequent minimum values of systolic and diastolic pressure differed between groups (Table 3). The minimum systolic pressure was higher in the OCT group than in the HF group (U = 822, p < .05) and tended to be higher in the OCT group than in the CABG group (U = 1451, p = .05). Minimum diastolic pressure was greater in both the OCT and HF groups than in the CABG group (U values = 1230 and 904, respectively; p values < .01).
Recovery From the Stroop Test
Minute-by-Minute Means.
Unadjusted data are shown in Figure 2. After adjustment for covariance with the measurements made at the end of the Stroop test, there was a main effect of patient group on the recovery of heart rate and diastolic pressure (Table 2 and Figure 3). Heart rate was highest in the OCT group, lower in the HF group (t = 2.89, p < .01), and lowest in the CABG group (t = 3.30 cf HF, p < .01). Diastolic pressure was also higher in the OCT group than in the CABG or HF group (t values = 2.89 and 2.63, respectively; p values < .01).
Significant interactions of group by time (Table 2) indicated that differences between the OCT and CABG groups on every variable increased as recovery proceeded, as did differences between the OCT and HF groups in heart rate and diastolic pressure (Figure 3). Accordingly, interactions of group by linear trend of time were significant for each variable (Table 2). Heart rate recovered less quickly in the OCT group than in the CABG group (t = 3.27, p < .01), but the comparison with the HF group did not reach significance. Systolic and diastolic pressure recovered less quickly in both the OCT and HF groups than the CABG group (systolic: minimum t value= 3.31, diastolic: minimum t value = 2.21, p values < .01 and .05, respectively)
Smoothed Data.
After adjusting for the maximum value reached during the Stroop test, groups differed in the minimum values of heart rate and systolic and diastolic pressure (Table 3). The minimum heart rate was higher in the OCT group than in the HF group (U = 500, p < .001) or CABG group (U = 486, p < .001). Minimum adjusted systolic and diastolic pressure were also higher in the OCT group than in the HF group (systolic: U = 822, p < .05; diastolic: U = 734, p < .01) and were higher, or tended to be higher, in the OCT than in the CABG group (systolic: U = 1451, p = .05; diastolic: U = 1023, p < .001).
Drug Effects
The drug regimen showed no specific association with hemodynamic responses to the Stroop test. Therefore, the results of the analyses of drug effects are only summarized here. Patients taking one or more of the cardiac agents had higher systolic and diastolic pressure at baseline and during recovery than those who did not, but there was no difference during the Stroop test. Differences associated with other types of drug were isolated, showing no clear pattern across the testing procedure. Patients taking cholesterol-lowering drugs had lower systolic and diastolic pressure at the end of the recovery period. Patients taking diuretics had lower heart rate at the start of Stroop test and lower systolic pressure at the end of recovery. Warfarin was associated with lower heart rate during the Stroop test. Allopurinol was associated with lower heart rate at the start of the Stroop test but higher heart rate at the end of the test. Finally, aspirin was associated with reduced heart rate during the Stroop test in the CABG group only.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
The lower systolic pressure in the HF group compared with that in the other groups was consistent with their clinical status. The similar systolic pressure in OCT and CABG patients during the initial relaxation phase suggests similar left ventricular function in these patients. This is to be expected: There is little direct innervation of the ventricular myocardium by the vagus, so its loss after OCT will have little, if any, impact on basal left ventricular function. Also, in this resting state, sympathectomy should have no effect. The high diastolic pressure in OCT, reported previously in a comparison with healthy subjects (9), could be due to hypertension associated with cyclosporin treatment (23). The finding that baseline diastolic blood pressure was lowest in HF patients illustrates their impaired cardiovascular function.
An important feature of this study was that we performed parallel measurements of the emotional response to the Stroop procedure. The similarity in psychological responses in the three patient groups indicates that their perception of the "stressfulness" of this procedure did not differ. This rules out differences in the central processing of stress as an explanation for any differences in their hemodynamic responses. Differences did emerge from the adjusted hemodynamic variables. Although there was an increase in all hemodynamic variables, in all groups of patients, during the Stroop (described below), OCT patients were generally less responsive to this stressor and adapted to it more slowly than the other groups, as was predicted. Moreover, they resembled HF patients more closely than CABG patients.
The smaller heart rate and systolic blood pressure responses to the stressor in OCT patients in comparison with CABG patients is consistent with disrupted sympathetic function. It is also consistent with the assumption, albeit controversial, that there is negligible reinnervation of a transplanted heart (24–28). Even if there is some reinnervation, it cannot be assumed that this is paralleled by any functional recovery (29). The smaller heart rate and systolic pressure responses in HF patients indicate that despite an intact cardiac innervation, these patients are impaired in responding to the Stroop test. However, a functional deficit of cardiac innervation would not explain why diastolic blood pressure was lower in the OCT group than in the CABG group. In fact, an increased diastolic pressure response might normally have been expected in the OCT group because of the hemodynamic adjustments needed to maintain mean arterial pressure. Because there is no such increase, it is possible that the hemodynamic response relies on hormonal secretion of adrenaline. Because adrenaline causes vasodilatation by activating vascular ß2-adrenoceptors, this might mask an underlying neurone-induced vasoconstriction.
In all patient groups, as in healthy subjects (8), the hemodynamic response reached its maximum during the first 1 to 2 minutes of the Stroop test, after which it slowly declined toward baseline. Despite their smaller stress response, this adaptation to the Stroop test was slower in both the OCT and HF groups than in the CABG group. This finding is further evidence of an exaggerated hormonal response in both the OCT and HF groups, which would dissipate more slowly than a neuronal one.
Although the minute-by-minute changes distinguished OCT patients from CABG patients, they did not clearly distinguish them from HF patients. Slight evidence of a distinction was evident in the more rapid changes detected by the smoothed data. OCT patients achieved a lower maximum rate of increase of heart rate than either HF or CABG patients. A reliance on hormonal catecholamines by OCT patients to respond to the Stroop test could explain this. In adapting to the Stroop test, systolic pressure reached a lower minimum in the OCT group than in the HF or CABG group, which is also consistent with the actions of hormonal catecholamines. The lower maximum diastolic pressure is difficult to explain because the sympathetic innervation of the vasculature should be unaffected by the transplantation. It is possible that this again reflects an adrenaline-induced vasodilatation.
During the post-Stroop (recovery) phase, slower recovery of each hemodynamic variable distinguished the OCT from the CABG group, although not from the HF group. Nevertheless, the heart rate and diastolic pressure were again both highest in the OCT group, even in comparison with the HF group, and the momentary values seen in the smoothed data once again confirmed poorer recovery in the OCT group than in the HF or CABG group. For the heart rate response, this slow recovery would be consistent with greater reliance on a hormonal outflow in the OCT patients that persists into the relaxation period. The reason why the high diastolic pressure should persist in these OCT patients during this recovery phase is harder to explain. Withdrawal of the adrenaline-induced vasodilation unmasking a neurogenic vasoconstriction is a possible explanation.
In summary, the absence of any differences in emotional responses to the Stroop rules out differences in the central processing of stress as an explanation for these patients’ different hemodynamic responses. Correspondingly, changes in the hemodynamic response to this stressor do not affect patients’ emotional responses. Within the OCT group, the role of the long-term deconditioning effect of heart failure that persists after surgery remains to be elucidated, as does the influence of underlying peripheral vascular disease. However, the HF group controlled for both these factors, which therefore cannot explain the differences, particularly in recovery from the Stroop test when the contrast with HF was clearest. Similarly, the contrast with HF patients cannot be attributed to the effects of cardiac disease that led to transplantation because this is shared by both groups. Because the pattern of a blunted response and slower adaptation in OCT patients contrasts with that in CABG patients, it cannot be attributed to the experience of major heart surgery. None of the drugs that we examined was associated with hemodynamic responses to the Stroop test, but the possibility that drug regimens and other aspects of treatment contributed to the results cannot be excluded.
By contrast, the main features of the disrupted responses to a psychological stressor after transplantation were predicted as a consequence of the functional deficit in the innervation of the heart and a greater reliance on a hormonal (adrenal) response. Transplanted patients’ responses to stress more closely resembled those of HF patients than CABG patients, emphasizing that despite their physical improvement, compromised psychophysiological functioning continues after this surgery. Further work is needed to explore the clinical implications of these findings. We previously showed that everyday physical symptoms contribute to continuing dysfunction after transplantation (30). Changed hemodynamic responses, if they occur in relation to stressors in normal life, might be one source of such symptoms. Successful rehabilitation might therefore require psychological intervention to reduce patients’ concern about them.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Received for publication April 26, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  S. Zipfel, A. Schneider, B. Wild, B. Lowe, J. Junger, M. Haass, F.-U. Sack, G. Bergmann, and W. Herzog Effect of Depressive Symptoms on Survival After Heart Transplantation Psychosom Med, September 1, 2002; 64(5): 740 - 747. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
