Anesth Analg 2000;91:329-336
© 2000 International Anesthesia Research Society
CRITICAL CARE AND TRAUMA
Could Heart Rate Variability Analysis Become an Early Predictor of Imminent Brain Death? A Pilot Study
Thierry Rapenne, MD*,
Daniel Moreau, PhD ,
François Lenfant, MD*,
Vincent Boggio, MD,
Yves Cottin, MD, PhD , and
Marc Freysz, MD, PhD*
*Département d’Anesthésie-Réanimation, Hôpital Général, CHU Dijon;
Centre d’Explorations Fonctionnelles; and
Centre de Cardiologie, Hôpital Le Bocage, CHU Dijon, Université de Bourgogne, Dijon, France
Address correspondence and reprint requests to Marc Freysz, MD, PhD, Département d’Anesthésie-Réanimation, Hôpital Général, 3, rue du Faubourg Raines, 21033 Dijon Cedex, France. Address e-mail to Marc.freyzz{at}CHU.fr
 |
Abstract
|
|---|
Physiology of brain death is characterized by major disturbances of autonomic nervous system (ANS) activity which can lead to graft dysfunction. These findings exhibit the importance of early diagnosis of brain death to improve transplantation outcome. The aim of this prospective study was to assess whether heart rate variability (HRV) analysis, a noninvasive method to investigate ANS activity in comatose patients, could achieve this goal. A total of 14 brain-injured patients were included in the study as soon as they exhibited the clinical signs of imminent brain death. The electrocardiogram was then recorded from two leads with a Holter digital monitor. The clinical diagnosis of brain death was considered after an autonomic storm had occurred. HRV was assessed from 6 h before to 6 h after brain death in both time domain and spectral analysis, estimating either global ANS activity (index of variability, total power), parasympathetic activity (percentage of  of R-R interval >50 ms, root mean square for successive interval differences, LnHF) or sympathetic activity (LnLF). Hourly averages of these variables were compared by using one-way analysis of variance. To assess whether HRV could per se diagnose brain death, receiver operating characteristic curves were generated for total power, root mean square for successive interval differences, and LnHF. We observed, for 6 h before brain death, a progressive extinction of the influence of the ANS on cardiovascular regulation. There was no activity in the two components of the ANS as soon as brain death occurred. HRV analysis appeared to be a very sensitive but a less specific method of diagnosing brain death.
Implications: A total of 14 brain-injured patients with the clinical criteria of imminent brain death were enrolled for electrocardiogram recording and heart rate variability analysis (a noninvasive method to investigate autonomic nervous system activity). For 6 h before brain death, we observed a progressive extinction of autonomic nervous system activity which was not present as soon as brain death was clinically evoked.
|
Introduction
|
|---|
In the past few years, fewer organs have been available for explantation, although the number of brain-dead patients has remained stable. Furthermore, approximately 20% of heart transplantations fail, many of them attributable to graft dysfunction. This dysfunction often results from the physiologic changes occurring in the heart before and during brain death (myocardial damage and electrocardiographic and hemodynamic changes) (1–3). These changes result mainly from the physiologic changes of the autonomic nervous system (ANS) activity during the autonomic storm preceding brain death and during the hemodynamic instability which follows. Observation of these changes may help early diagnosis of brain death to improve transplantation outcome.
Instantaneous heart rate is not steady but rather, demonstrates continuous fluctuations. As these oscillations in instantaneous heart rate depend largely on interactions between sympathetic and parasympathetic efferent activities, heart rate variability (HRV) analysis has been widely used as a measure of activity in the two components of ANS (4–6). After myocardial infarction, the relative risk of mortality was reported to be 5.3 times higher in patients with decreased HRV (7). In patients with diabetes, HRV analysis is a sensitive method for early detection of autonomic neuropathy (8). After brainstem ischemia and necrosis, the brain-heart connections are definitively disrupted. This disconnection should logically lead to early decrease of HRV. We, therefore, postulated that HRV analysis could be an early criterion of brain death and could, in the future, accelerate further brain death certification, clinical assessment, and other validating proceedings, such as electroencephalography (EEG) or cerebral angiography. In Europe, there are large differences among countries regarding brain death certification (9). Our aim was, therefore, to continuously assess HRV of comatose patients before brain death.
|
Methods
|
|---|
After approval by the institutional ethical committee (Comité de Protection des Personnes dans la Recherche Biomédicale de Bourgogne), severely brain-injured patients were included in this prospective study as soon as they exhibited the clinical criteria of imminent brain death, determined by an algorithm, at the latest time point when an acute hyperdynamic state reflecting autonomic storm occurred. Patients were admitted to a trauma intensive care unit and managed according to a specific protocol of resuscitation (Figure 1). The study interfered in no way with local routine management. The criteria of imminent brain death were strongly elevated and increasing intracranial pressure, impossible to lower despite optimal treatment, and lack of brainstem reflex. Brain death diagnosis was confirmed by an apnea test, two silent EEGs, or a blocked cerebral blood flow in carotid and vertebral angiography (10).
View larger version (19K):
[in this window]
[in a new window]
|
Figure 1. Algorithm of recruitment and management of the studied patients. GCS = Glasgow Coma Scale score.
|
|
We excluded patients with known cardiac disease, arrhythmia, and specific medications (atropine, ß-blockers), and with clinical factors which could alter heart rate, such as hypoxia defined by <80 mm Hg PaO2, <100 mm Hg systolic blood pressure, anemia defined by <100 g/L hemoglobin, spinal cord trauma, and diabetes mellitus. The following data were recorded in all patients: age, etiology of brain injury, Simplified Acute Physiology Score II, and the score of the initial computerized tomography scan using the classification of the Traumatic Coma Data Bank.
Before the autonomic storm, patients were managed according to a written protocol of resuscitation whose goal was to obtain an optimum cerebral perfusion pressure (CPP) >70 mm Hg (11). The heart rate, invasive arterial pressure, and central venous pressure were used as hemodynamic variables. Intracranial fiber optics were used to monitor intracranial pressure (ICP). All of these values were recorded continuously from bedside monitors (Merlin; Hewlett-Packard, Palo Alto, CA). Patients were kept in the supine position with 15° inclination. In patients with ICP greater than 25 mm Hg, jugular venous oxygen saturation was monitored. All patients were tracheally intubated and artificially ventilated; neurosedative drugs (flunitrazepam, fentanyl, and, if necessary, propofol or thiopental) were administered to facilitate ventilation, to decrease ICP, and to decrease cerebral oxygen consumption. Ventilation was adjusted to maintain PaCO2 around 35 mm Hg. The aim of fluid management was to maintain euvolemia and normonatremia. To maintain >70 mm Hg CPP, ICP was decreased when possible and blood pressure was increased. Fluids and vasopressors were used to increase blood pressure. Neurosedative therapy, hyperventilation, mannitol, or hypertonic saline solution and cerebrospinal fluid external drainage were used to decrease ICP. Clinical diagnosis of brain death was considered after an autonomic storm (acute hyperdynamic state) had occurred or when clinical signs of brain death were present. After this, perfusions of neurosedative drugs were stopped. After infusion of flumazenil and naloxone, an apnea test was done. If positive (no breathing despite normoxia and >60 mm Hg PaCO2), brain death was confirmed with carotid and vertebral angiographies or two EEGs in accordance with French legislation (9). Fluid administration and vasopressor infusion were continued to maintain hemodynamic variables in the normal ranges. If no organ could be recovered, fluid and catecholamine infusions were stopped (more than 6 h after the autonomic storm).
Long-term electrocardiograms were recorded continuously from two leads with a Holter digital monitor (Synesis; Ela Medical, Le Plessis Robinson, France) as soon as clinical signs of imminent brain death appeared. HRV analysis per se was performed from 6 h before to 6 h after autonomic storm. The analysis of each recording was performed on the Elatec system (Ela Medical). After classification of QRS morphology, the longest and the shortest R-R intervals were manually confirmed until no QRS was mislabeled as either an artifact or an extra systole. Only cycles in which the QRS had normal morphologic characteristics were used for HRV analysis. HRV was assessed in two ways: first, by time domain analysis based on statistical operations on R-R intervals and second, by spectral analysis. Three indices of HRV were distinguished in time domain analysis. Two short-term variability indices, a root mean square for successive interval differences (rMSSD) and percentage of of R-R interval >50 ms (pNN50), are mediated by the parasympathetic nervous system. A third index, the index of variability, reflects global HRV and is mediated by the two components of the ANS. Spectral analysis decomposes the series of sequential R-R intervals into peaks of different amplitudes and frequencies. The beat-to-beat fluctuations were transformed into the frequency domain by using a fast Fourier transformation. The heart rate spectra were quantified by determining areas of integral spectra and areas under two components: a high frequency (HF) component, from 0.15 to 0.40 Hz, used as a measure of parasympathetic-mediated respiratory sinus arrhythmia, and a low frequency (LF) component, from 0.04 to 0.15 Hz, used to quantify ß-sympathetic mediated heart rate fluctuation. To confirm a normal distribution, log power and log total power (TP) of these two components were calculated. Values of HRV variables in healthy subjects are shown in Table 1.
A report on hourly averages of systolic blood pressure (SBP), diastolic blood pressure (DBP), and mean blood pressure (MBP), ICP, CPP, index of variability, pNN50, rMMSSD, log TP spectrum, log LF power, log HF power and the rate of the vasoactive and neurosedative drug infusions was analyzed for each patient. All data were expressed as mean ± SE to mean (SEM). Hourly averages of each index were analyzed and compared by using one-way of analysis of variance followed by Student-Newman-Keuls test. P < 0.05 was considered significant.
Receiver operating characteristic (ROC) curves were generated for TP, log of high frequency power, and rMSSD, varying the discriminating threshold of each variable. The area under each ROC curves was calculated. Values for each area can be between 0 and 1. A value of 0.5 indicates that the screening measure is not better than chance, whereas a value of 1 implies perfect performance to diagnose brain death. The optimal threshold value (that maximizes the sum of the sensitivity and specificity) was also calculated for each variable. Because some authors denied the reliability of HRV as a sympathetic index and because of the spinal cord origin of the sympathetic nervous system, only variables reflecting either global HRV or more specifically parasympathetic tone were affected by ROC curve construction (12).
|
Results
|
|---|
Fourteen consecutive patients (12 men, 2 women) were included in this study. Because of the unavailability of the Holter digital monitor, one patient fulfilling all inclusion criteria was excluded. Five patients had monitoring 6 h before the autonomic storm (H-6), two patients 5 h before (H-5), one patient 4 h before (H-4), one patient 2 h before (H-2), one patient 1 h before (H-1), and finally three patients when this storm occurred (H0). A multiorgan donation (including heart donation) was performed in seven patients. Organs were not removed from seven patients because of family refusal. No further information is available about the viability of the grafts.
Mean age was 36.9 ± 15.9 yr and the mean Simplified Acute Physiology Score II was 42.9 ± 6.9. Ten patients were involved in road accidents. The other admitting diagnoses were gunshot wounds to the head (n = 2), fall (n = 1), and cerebral anoxia (n = 1). Systemic injuries associated were chest trauma (n = 3), bone trauma (n = 4), and pelvis trauma (n = 1). On initial cerebral computerized tomography scan, most patients (11 of 14) presented either absent or compressed cisternae, midline shift >5 mm, or mixed density lesion >25 mL (surgically evacuated or not).
The hourly mean rate of the vasoactive and neurosedative drug infusions are shown in Table 2. No significant difference was noticed between the hourly averages of any therapeutic concentration levels except for thiopental infusion. All administration of this product was stopped after the autonomic storm occurred (as stipulated in the therapeutic protocol used) and statistical difference was noticed between the two following stages, before and after brain death.
Systemic and intracranial hemodynamic changes are shown in Figure 2. Hemodynamic variables remained stable during the first 6 h of recording, except just before brain death (during autonomic storm). At this moment, patients experienced a nonsignificant increase of blood pressure associated with higher levels of ICP. Brain death was followed by a transient significant increase in heart rate. More than 2 h after the autonomic storm, this increase was no longer significant, however, heart rate never returned to prior values. After brain death, all systemic hemodynamic variables (SBP, MBP, DBP) declined. This decrease was significant 2 h after brain death (H+2).
View larger version (28K):
[in this window]
[in a new window]
|
Figure 2. Hourly heart rate (HR, bpm), systolic blood pressure (SBP, mm Hg), diastolic blood pressure (DBP, mm Hg), and mean blood pressure (MBP, mm Hg) over the 6-h period before (H-6 to H-1) and 6-h period after (H+1 to H+6) brain death. Data are presented as mean ± SEM and were compared by using one-way analysis of variance followed by Student-Newman-Keuls test (P < 0.05).
|
|
During the first 6 h before autonomic storm, ICP increased progressively to reach a peak value of >100 mm Hg. This increase became significant just before brain death, when a Cushing reflex occurred (Figure 3). Then, ICP decreased and settled at a steady-state level of 90 mm Hg. CPP showed a continuous decrease from the first hours of recording. This decrease was significant after H-2. The autonomic storm, and thus, brain death, occurred when CPP was in a low positive range, approximately 20 mm Hg. After brain death, CPP was near to 0 mm Hg and often became negative (Figure 3).
View larger version (28K):
[in this window]
[in a new window]
|
Figure 3. Hourly intracranial pressure (ICP) and cerebral perfusion pressure (CPP) over the 6-h period before (H-6 to H-1) and 6-h period after (H+1 to H+6) brain death. Data are presented as mean ± SEM and were compared by using one-way analysis of variance followed by Student-Newman-Keuls test (P < 0.05).
|
|
Results of time domain and spectral analyses of HRV are shown in Figures 4 and 5. After brain death, the overall HRV was small. The mean of the two variables which reflect global HRV (index of variability and TP) were very low (0.6% and 20 vs 1.8% and 700 ms2 6 h before the autonomic storm, respectively). Both the HRV analysis approaches showed a significant collapse of both parasympathetic (pNN50, rMSSD, LnHF) and sympathetic (LnLF) activities after autonomic storm. All of these variables reached a very low steady-state level after H0.
View larger version (26K):
[in this window]
[in a new window]
|
Figure 4. Results of the statistical analysis of HRV. Hourly IV (%), rMSSD (ms), and pNN50 (%) over the 6-h period before (H-6 to H-1) and 6-h period after (H+1 to H+6) brain death. Data are presented as mean ± SEM and were compared by using one-way analysis of variance followed by Student-Newman-Keuls test (P < 0.05). IV = index of variability; rMSSD = root mean square for successive interval difference; pNN50 = percentage of of R-R interval for >50 ms.
|
|
View larger version (35K):
[in this window]
[in a new window]
|
Figure 5. Results of the spectral analysis of HRV. Hourly total power (ms2), logarithm of low frequency power (LF) and logarithm of high frequency power (HF) over the 6-h period before (H-6 to H-1) and 6-h period after (H+1 to H+6) brain death. Data are presented as mean ± SEM and were compared by using one-way analysis of variance followed by Student-Newman-Keuls test (P < 0.05). Ln = logarithm to the natural base.
|
|
Moreover, patients experienced, from 6 h before brain death, a progressive and significant decrease in the influence of the ANS on cardiovascular regulation. This gradual decrease concerned the two mean components of the ANS. Decrease in hourly values became significant after H-3 in time domain analysis and after H-4 in spectral analysis. Finally, and just before brain death, this study showed a little reinforcement of both sympathetic and vagal influence when the autonomic storm occurred.
The overall performance of rMSSD, LnHF, and TP to predict the imminent evolution to brain death was evaluated by constructing ROC curves (Figure 6). The area under the ROC curves was 0.32, 0.37, and 0.37, respectively. The optimal thresholds were rMSSD, 9.94 ms (sensitivity 0.94, specificity 0.24); LnHF, 2.10 (sensitivity 0.94, specificity 0.27) and TP, 83.62 (sensitivity 0.90, specificity 0.27).
View larger version (16K):
[in this window]
[in a new window]
|
Figure 6. Total power, rMSSD and LnHF receiver operating characteristic curves. rMSSD = root mean square for successive interval difference, LnHF = log of low frequency power.
|
|
|
Discussion
|
|---|
Our aim was to assess, continuously, HRV of patients just before brain death. Our first problem was to determine the exact time of the occurrence of brain death. HRV analysis was, indeed, performed at least six hours after this presumed occurrence of brain death. As it appears in practice impossible to determine this phenomenon with exact precision, brain death was considered as effective when all of the following criteria were fulfilled: deep coma (Glasgow Coma Scale score of 3), lack of brainstem reflexes (in routine the photo-pupil reflex; the nasopalpebral, corneal, vertical and oculocephalic reflex; and the oculocardiac reflex) and the occurrence of an acute hyperdynamic state (reflecting an autonomic storm). The existence of this hyperdynamicstate taking place during the development of brain death is now well known (1,3,13–15,17–20).1 Although ICP and CPP were not used with this end in view, they corroborated our correct determination of the occurrence of brain death. Indeed, the acute hemodynamic changes we observed were concomitant first with the highest levels of ICP and this phenomenon occurred when CPP was in a low positive range—approximately 20 mm Hg. Van Loon et al. (14) investigated catecholamine response to a gradual increase of ICP. When CPP decreased to a range of 20 to 30 mm Hg, circulating epinephrine and norepinephrine levels increased rapidly. In the upper range of CPP, circulating catecholamine levels were normal.
As soon as the diagnosis of brain death was clinically suggested, the HRV analysis demonstrated a lack of control of the two components of the ANS on cardiovascular regulation. The variables reflecting either the sympathetic tone (LnLF) the parasympathetic efferent activity (pNN50, rMSSD, LnHF), or the overall HRV (index of variability, TP) approached zero levels, with lack of interindividual variability. These data were in agreement with the results of other studies (16, 22–24). Mean rMSSD was 7.7 msec in 12 brain-dead children compared with 30 to 90 msec for normal healthy children (24). Many investigators have compared heart rate power spectra in the resting brain-dead, compared with the resting vegetative, patients. LF power (16,23,24), and HF power spectra (24) which reflects parasympathetic tone, significantly decreased or disappeared after brain death. The short-term modulation of the cardiovascular system is mainly controlled by the ANS (25). Neuroanatomy facilitates recognition and analysis of the neurocardiologic links. The heart receives neural input both from parasympathetic ganglia located in medulla oblongata and from the sympathetic tract located in the intermediolateral gray column of the spinal cord (C-8 to L-2). These neural regulatory outflows are controlled by various regions in the nervous system. Most of the higher centers (especially the cortex and the amygdaloid complex) provide descending efferent connections with all lower levels. In brain-dead patients, all of these connections are disrupted and destroyed. Brain death results in complete cessation of normal variations of the autonomic cardiovascular centers. The catecholamine circulating level decreases after brain autonomic storm (16). In the current study, a very small LF power spectrum could be found in these patients; free from regulation by the higher centers, the sympathetic nerves of the spinal cord continue to generate small autonomic impulses to control vasomotor tone. In the same way, a very small HF power spectrum could be found without clear explanation (could HF peaks also reflect parasympathetic nervous activity?).
According to our results and previous reports (16, 22–24), HRV analysis may prove to be a useful adjunct in the determination of brain death. HRV collapsed as soon as brain death was clinically suggested. This could become a very early promising indicator of imminent brain death. However, could HRV analysis, easily monitored at the bedside, become an independent criterion of brain death? Referring to the area under rMSSD, LnHF, and TP ROC curves, one can easily doubt this. Our results show that HRV performs poorly in the diagnosis of brain death. HRV analysis was marked by a very good sensitivity associated to a very low specificity. An explanation for this very low specificity can be detected in our protocol management of severely head-injured patients. Most of these patients were sedated before the occurrence of brain death. Benzodiazepines, thiopental, and propofol are now well known to decrease HRV (26). Those therapies thus, certainly decreased the specificity of HRV analysis for diagnosing brain death. Nevertheless, HRV could be promising when the diagnosis of brain death is suspected in patients free from any sedation (e.g., after cardiac arrest leading to cerebral anoxia). Other larger studies should be conducted to test how HRV analysis in unsedated patients could accelerate further validated confirmative diagnosis of brain death, and so decrease graft dysfunction. At least, HRV could reduce the long uncertainty and suffering for the patient’s family and frustration for the intensive care unit team.
Until now, there have been no reports of continuous analysis of both sympathetic and parasympathetic activities in patients before brain death. Previous experimental studies compared repetitive circulating or interstitial myocardial concentrations of catecholamines before and after brain death (27). Other clinical studies analyzed HRV in severely brain-damaged and brain-dead patients. Only 500 or less (often 256) representative beat-to-beat intervals were measured per record (16,22–24).
With the following restriction that not all of our patients were studied for 6 h before the presumed occurrence of brain death, both in the time domain and spectral analyses, the influence of the two components of the ANS on cardiovascular control showed an early, gradual, and irremediable decrease beginning 6 h before brain death and were not present after the autonomic storm had occurred. Referring to literature data, the cause of this decrease remains unknown and few hypotheses can explain it. One possibility could be the influence of the hourly evolution of the rate of both sedative and vasopressive drug infusions on HRV analysis. In the current study, no significant differences between these hourly values were noted, except for thiopental (Table 2). However thiopental infusion remained stable before brain death. In addition, the first hypothesis cannot explain the gradual decrease in sympathetic and parasympathetic activities. A second possibility might be the progressive anoxia of both of the medulla oblongata and the higher regulating centers of the ANS. The CPP evolution curve strengthened this hypothesis. In the current study, the general evolution of both CPP, on the one hand, and rMSSD, pNN50, Ln HF and LnLF on the other, was remarkably similar.
In summary, our study showed a progressive and significant decrease, from 6 h before brain death, of both sympathetic and parasympathetic outflows to the heart. After the autonomic storm occurred, sympathetic and vagal tone were not present. Accordingly, HRV analysis could indicate the imminence of brain death and could be used, after further validation, as one of the early indicators of brain death and its diagnosis, especially in patients free from neurosedative therapy.
|
Appendix 1. Concept of HRV
|
|---|
It has long been recognized that the instantaneous heart rate is not steady, but fluctuates from beat to beat. "Heart rate variability" is the conventionally accepted term used to describe the oscillations in the interval between consecutive instantaneous intervals. Such oscillations in instantaneous heart rate depend largely on the interactions between sympathetic and parasympathetic efferent impulses to the heart. Thus, analysis of HRV has become the most commonly used measure of the cardiovascular autonomic regulatory system, reflecting both sympathetic and parasympathetic functions. The variations in heart rate may be evaluated by several approaches. The most frequently used are statistical (time domain) and spectral (frequency domain) methods.
Time Domain Method
After continuous electrocardiogram recording, each QRS complex is detected and all intervals between adjacent QRS complexes resulting from sinus node depolarization are analyzed. Few variables, resulting from long recordings (traditionally 24 h) or smaller ones can be calculated (Tables 1 and 3). The standard deviation of all R-R intervals and the standard deviation of the averages of R-R intervals in all 5-min recordings reflect all components responsible for HRV in the period of recording. rMSSD and pNN50 reflect parasympathetic heart rate modulation.
Frequency Domain Method
Spectral analysis decomposes HRV into a sum of sinusoidal functions of different amplitudes and frequencies. The methods most commonly used for spectral analysis are based on the discrete Fourier transform and autoregressive modeling. Most investigators prefer this method rather than the statistical approach. The main advantage of spectral approach is the ability to obtain information about, not only the amount of variability, but also about the oscillation frequencies (number of heart rate fluctuation per second). Three main spectral components are generally distinguished: very low frequency component (VLF) <0.04 Hz, low frequency component (LF) 0.04–0,15 Hz, and high frequency component (HF) 0.15–0.40 Hz. HF reflects respiratory modulations of heart rate and is widely considered to be an index of parasympathetic efferent activity. LF assesses the baroreflex-related heart rate fluctuations and, in this way, assesses both sympathetic and vagal efferent activities. The significance of the VLF component is not yet clearly understood. It has been suggested that it reflects thermoregulation-related heart rate modulations.
|
Acknowledgments
|
|---|
Supported, in part, by a grant from the Clinical Research Committee, Centre Hospitalier Universitaire, Dijon, France.
The authors thank Dr. Paul Walker and Dr. Magali Vernet for a critical reading of this manuscript. The authors are indebted to Ms. Bernadette Daumas for her critical assistance.
|
Footnotes
|
|---|
1 Goarin JP, Cohen S, Jacquens Y, et al. Left ventricular function in brain dead donors: assessment using transesophageal echocardiography (TEE) [abstract]. Anesthesiology 1990;73:A84. 
|
References
|
|---|
-
Novitzky D, Rose AG, Cooper DKC. Injury of myocardial conduction tissue and coronary artery smooth muscle following brain death in the baboon. Transplantation 1988;45:964–6.[Web of Science][Medline]
-
White M, Wiechmann RJ, Roden RL, et al. Cardiac beta adrenergic neuroeffector systems in acute myocardial dysfunction related to brain injury. Circulation 1995;92:2183–9.[Abstract/Free Full Text]
-
Cooper DKC, Novitzky D, Wicomb WN. The physiopathological effect of brain death on potential donor organs, with particular reference to the heart. Ann Roy Coll Surg Engl 1989;7:261–6.
-
Akselrod S, Gordon D, Andrew Hubel F, et al. Power spectrum analysis of heart rate fluctuation: a quantitative probe of beat-to beat cardiovascular control. Science 1981;213:220–2.[Abstract/Free Full Text]
-
Malik M. Heart rate variability: standard of measurement, physiological interpretation and clinical use. Circulation 1996;93:1043–65.[Free Full Text]
-
Malliani A, Pagani M, Lombardi F, Cerutti S. Cardiovascular neural regulation explored in the frequency domain. Circulation 1991;84:482–92.[Abstract/Free Full Text]
-
Kleiger RE, Miller JP, Bigger JT, et al. Decreased heart rate variability and its association with increased mortality after acute myocardial infarction. Am J Cardiol 1987;59:256–62.[Web of Science][Medline]
-
Weise F, Heydenreich F, Runge U. Heart rate fluctuations in diabetic patients with cardiac vagal dysfunction: a spectral analysis. Diabet Med 1988;5:324–7.[Web of Science][Medline]
-
Haupt WF, Rudolf J. European brain death codes: a comparison of national guidelines. J Neurol 1999;246:432–7.[Web of Science][Medline]
-
Louvier N, Combes JC, Nicolas F, et al. Cerebral angiography must have medicolegal value for brain death confirmation in France. Transplant Proc 1996;28:377.[Web of Science][Medline]
-
Rapenne T, Lenfant F, N'Guyen Kim, et al. Facteurs predictifs de mortalité à court terme chez le patient traumatisé cranien grave. Presse Med 1997;26:1661–5.
-
Hopf HB, Skyschally A, Heusch G, Peters J. Low-frequency spectral power of heart rate variability is not a specific marker of cardiac sympathetic modulation. Anesthesiology 1995;82:609–19.[Web of Science][Medline]
-
Rosner MJ, Newsome HH, Becker DP. Mechanical brain injury: The sympathoadrenal response. J Neurosurg 1984;61:76–86.[Web of Science][Medline]
-
Van Loon J, Shivalkar B, Plets C, et al. Catecholamine response to a gradual increase of intracranial pressure. J Neurosurg 1993;79:705–9.[Web of Science][Medline]
-
Clifton GL, Robertson CS, Kyper K, et al. Cardiovascular response to severe head injury. J Neurosurg 1983;59:447–54.[Web of Science][Medline]
-
Golstein B, Kempski MH, De King D, et al. Autonomic control of heart rate after brain injury in children. Crit Care Med 1996;24:234–40.[Web of Science][Medline]
-
Mertes PM, Carteaux JP, Jaboin Y, et al. Estimation of myocardial interstitial norepinephrine release after brain death by using cardiac microdialysis. Transplantation 1994;57:371–7.[Web of Science][Medline]
-
Evans DE, Alter WA, Shatsky SA, Gunby EN. Cardiac arrhythmias resulting from experimental head injury. J Neurosurg 1976;45:609–16.[Web of Science][Medline]
-
Connor RCR. Myocardial damage secondary to brain lesions. Am Heart J 1969;78:145–8.[Web of Science][Medline]
-
Rona G. Catecholamine cardiotoxicity. J Mol Cell Cardiol 1985;17:291–306.[Web of Science][Medline]
-
Deleted in proof.
-
Kero P, Antilla K, Ylitalo V, Valimaki I. Decreased heart rate variability in decerebration syndrome: quantitative clinical criterion of brain death? Pediatrics 1978;62:307–11.[Abstract/Free Full Text]
-
Kita Y, Ishise J, Yoshita Y, et al. Power spectral analysis of heart rate and arterial blood pressure oscillation in brain death patients. J Auton Nerv Syst 1993;44:101–7.[Web of Science][Medline]
-
Lacquaniti LG, Irone M, Barbacini S, et al. Heart rate variability and severe brain damage: preliminary data. Int J Clin Monit Comput 1993;10:181–5.[Web of Science][Medline]
-
Talman WT. Cardiovascular regulation and lesions of the central nervous system. Ann Neurol 1985;18:1–12.[Web of Science][Medline]
-
Howell SJ, Wanigasekera V, Young JD, et al. Effects of propofol and thiopentone and benzodiazepine premedication on heart rate variability measured by spectral analysis. Br J Anaesth 1995;74:168–73.[Abstract/Free Full Text]
-
Powner DJ, Hendrich A, Nyhuis A, Strate R. Changes in serum catecholamine levels in patients who are brain dead. J Heart Lung Transplant 1992;11:1046–5.[Web of Science][Medline]
Accepted for publication April 13, 2000.
|
Top
Abstract
Introduction
