- Email: [email protected]
Somatosensory evoked potentials following severe head injury: Loss of the thalamic potential with brain death
Journal of the Neurological Sciences, 1988, 87:255-263 Elsevier
255
JNS 03062
Somatosensory evoked potentials following severe head injury: loss of the thalamic potential with brain death A . M . Chancellor, R.W. Frith and N.A. Shaw Department of Clinical Neurophysiology, Auckland Hospital, Park Road, Auckland (New Zealand) (Received 19 April, 1988) (Revised, received 18 June, 1988) (Accepted 20 June, 1988)
SUMMARY
The thalamic component (P17) of the short-latency somatosensory evoked potential (SEP) was assessed to determine its usefulness in patients with severe head injury. Subjects were a group of patients admitted to the Auckland Hospital Critical Care Unit who subsequently died from head injury. In all instances where brain death was unequivocally established and a SEP recording made in close temporal proximity to the time of brain death the P17 potential was absent. When there was evidence of continuing brainstem activity and particularly where prolonged survival occurred following the last SEP recording the P17 potential remained intact bilaterally. This study shows that the presence or absence of the thalamic component of the short-latency SEP provides a reliable electrophysiological measure of bralnstem function in patients where brain death has been suspected.
Key words: Brain death; Head injury; Somatosensory evoked potential; Thalamic potential
Correspondence to: Dr. R.W. Frith, Neurologist and Clinical Neurophysiologist, Department of Clinical Neurophysiology,Auckland Hospital, Park Road, Auckland l, New Zealand. 0022-510X/88/$03.50 © 1988 Elsevier Science Publishers B.V. (Biomedical Division)
256 INTRODUCTION With the availability of advanced life support systems information is needed on how best to define brain death in the presence of continued cardiopulmonary function. The diagnosis of brain death is usually made by clinical examination although in some centres angiography or an isoelectric EEG in the absence of drug or hypothermic effects provides useful laboratory conf'trmation. The role of other neurophysiological tests, in particular sensory evoked potentials (SEPs), remains uncertain. Short-latency evoked potentials which arise in the primary sensory cortex or more caudal locations have an advantage over the EEG in remaining stable even when the patient has received sedative drugs (Cant and Shaw 1986). A number of studies have now examined the role of brainstem auditory evoked potentials (BAEPs) in the determination of brain death (Starr 1976; Goldie et al. 1981; Hall et al. 1985; Kaga et al. 1985; Garcia-Larrea et al. 1987). The principal waves of the BAEP arise within the 8th cranial nerve and brainstem auditory pathways and therefore these potentials represent a direct measure of brainstem activity. A major problem complicating the use of the BAEP in the diagnosis of brain death is that, for unequivocal interpretation, wave 1 needs to be preserved to indicate that the afferent volley is being propagated within the auditory nerve. Often, however, the entire BAEP complex is abolished. Under these circumstances no certain conclusions can be drawn about the integrity of brainstem structures as complete absence may indicate acute or pre-existing cochlear or auditory nerve damage (Goldie et al. 1981). In addition to the BAEP there have also been reports of short-latency somatosensory evoked potentials (SEPs) in brain death (Trojaborg and Jorgensen 1973; Goldie et al. 1981; Anziska and Cracco 1980; Belsh and Chokroverty 1987; Buchner et al. 1988). Usually, the short-latency SEP consists of the simultaneous recording of the cortical and cervical SEP and there is consensus that while the cortical SEP is invariably lost in brain death, the cervical SEP may be preserved. This pattern of abnormality does not, however, specify at which level within the central somatosensory pathways damage has occurred. Hence the absence of a cortical potential with a preserved cervical potential cannot be taken as indicating loss of brainstem function. SEP components thought to reflect brainstem function have been identified and their disappearance correlated with brain death (Belsh and Chokroverty 1987). The difficulty with such components is that they are not optimally recorded using a cephalic reference and are often ill-defined or not reliably recorded in normal subjects. The principal component of the cortical SEP is a negativity usually labelled N20 which is assumed to be generated within the primary somatosensory cortex (Cant and Shaw 1986). The cervical component appears to be a complex amalgam of activity arising in the cervical roots, dorsal columns, cuneate nuclei and the medial lemnisci. However, the major component of the cervical SEP, the N13 potential, arises within the interneurones of the dorsal horn (Desmedt and Cheron 1981). Following the downward slope of the major negative potential lies a positive trough which, according to its mean latency, has been labelled P17. It has been suggested that this positivity is a far-field reflection of activity generated in or near the thalamus (Shaw 1985). P17
257 probably reflects synaptic and post-synaptic activity from the ventroposterolateral nucleus of the thalamus. P17 is generally robust, stable and is easily recorded. If the potential does arise in the diencephalon it should be absent bilaterally in all patients with brain death. In addition its loss in sequential recordings may be a useful indicator of continuing rostro-caudal deterioration of cerebral function. SEPs recorded from head-injured patients may be divided into three major types (Fig. 1). Type 1 is a normal SEP with intact cortical and cervical wave forms and a normal central conduction time. The central conduction time (CCT) is calculated by subtracting the latency of the cervical potential from that of the cortical potential (Cant and Shaw 1986). Mean CCT is 5.6 msec in adults and a CCT greater than 7.1 msec is abnormal in this laboratory. Type 2 SEP consists of an absent cortical potential but with preservation of both the N13 and P17 components of the cervical potential. It has been shown previously that a type 2 SEP recorded from head-injured patients either unilaterally or bilaterally is associated with an unfavourable outcome (Cant et al. 1986). The type 3 SEP differs from type 2 in that there is loss of the P17 component with preservation of the N13 potential. The purpose of the present study was to investigate the relationship between the loss of the thalamic potential and brain death. We tested the hypothesis that a type 2 SEP was associated with clinical evidence of continuing brainstem function but that a type 3 SEP indicated brain death.
PATIENTS AND METHODS
Subjects were 36 severely head-injured patients who were admitted consecutively to the Department of Critical Care Medicine at Auckland Hospital and who subsequently died of their head injury. Mean survival time was 7 days (median 3 days, range 16 h to 75 days). There were 29 males and 7 females. Median age was 22 years (range 15-73 years). The initial Glasgow Coma Score was less than 8 in all patients. SEPs were recorded on at least 3 of the in'st 4 days foUowing head injury, if the patient survived, and at variable intervals thereafter. Twenty-five patients satisfied clinical criteria for brain death. The remainder had active support withdrawn following a clinical decision that the brain injury was unsurvivable and most died within a short period of time. Three patients remained in a persistent vegetative state for days to weeks prior to death. The diagnosis of brain death was made according to standard criteria based on guidelines of the President's Commission Report (Guidelines 1981). These were: (a) that the patient was not sedated, paralysed or hypothermic, (b) there was no motor activity in the limbs other than that due to spinal reflexes, (c) brainstem reflexes were absent, including, (d) apnoea in the presence of a Paco~ elevated to 60 mm Hg with maintenance of oxygenation. All patients had a postmortem examination which confumed that death was a consequence of head injury. SEPs were recorded using a Tracor-Northern 3 000 evoked response analyser. The median nerves at the wrists were stimulated independently on each side using square wave pulses of 0.15 msec duration presented at a rate of 4.1/sec. Recording
258 electrodes (platinum alloy needles) were inserted over the cervical spine (CII) and over the contralateral sensory cortex (C 4' or C3', 2 cm posterior to C a or C 3, international 10-20 system). A common mid-forehead reference (Fpz) was used. Cervical and cortical SEPs were recorded simultaneously. Analysis time was 40 msec with the first 4 msec following stimulus onset excluded. Sampling interval was 80 #sec. The bandpass of the recording system was 10 Hz to 3 kHz. Each set of SEPs was the average of 500 samples and all SEPs were replicated and superimposed in order to demonstrate reproducibility. Patients were divided into 2 groups according to whether the thalamic component (P17) of the cervical potential was absent or present in the fmal SEP recorded before death. Those in group A had absence of the P17 potential bilaterally in the final SEP recording before death. In group B, the P17 was preserved bilaterally in the final SEP performed prior to death. The 2 groups were compared to determine if loss of the P 17 potential correlated with the patient's clinical status at the time of the last recording.
RESULTS
Figure 1 illustrates the 3 types of short-latency SEPs recorded from patients with severe head injury. Figure 2 illustrates SEPs recorded on 3 consecutive days from one patient. On days 3 and 4 the 3 principal components, N13, P17 and N20, can be identified. However, the cortical potential is reduced in amplitude and there is an
TYPE NORMAL
t
TYPE SEP
LOSS
2
OF
PRESERVATION AND
TYPE N20 OF
LOSS P17
OF
3
N20
AND
PRESERVATION
OF
P17 Nt3
N13
N20
CORTICAL SEP
~
CSVCAL EP '~t' p 1 ? '~r P 1 7 _ ~ + t ~Jv 4 MS
Fig. 1. Examples of three types of short-latency somatosensory evoked potentials (SEPs) recorded from patients following severe head trauma. In type 1 the SEP is normal. The 3 principal components of the waveforms (N13, P17 and N20) are labelled. The central conduction time is 6.5 msec. Central conduction time cannot be calculated in either type 2 or type 3 SEPs because the cortical waveform has been lost. In type 2 SEPs the P17 is preserved while in type 3 it is absent.
259 D AY 3
J CORTICAL SEP
.
~
~
DAY 4
DAY 5
• N20 ~
~ ~
~
..~.~
( C 4' - Fpz )
N I 3 ,~,
( C ii-
N13
Fpz ) P17 _~+1 uv LEFT
MEDIAN
NERVE
STIMULATED
4 MS
Fig. 2. SEPs recorded on 3 consecutive days. On days 3 and 4 both N20 and P17 are preserved although the central conduction time is prolonged. On day 5 neither PIT nor N20 are present. The patient died approxL,nate]y 4 h after the day 5 recording.
abnormal prolongation in CCT to 8 msec. On the following day (day 5) both the cortical and thalamic components had disappeared bilaterally although N13 persisted. The patient died on the same day as the last recording.
Patient data Group A There were 11 patients in this group in whom the P17 was absent bilaterally in the last recording. In 9 of the group brain death was declared prior to, or occurred within 4 h after, the recording of bilaterally absent P17 and cortical potentials. In the other 2 patients a formal brain death assessment was not made until the following day (within 20 h of the recording) but the clinical notes suggest that brain death had occurred on the day of the recording. All patients in group A had isoelectric EEGs recorded at the time of the final SEP. In 2 patients the first and only recording showed a low amplitude N 13 with absent P17 and cortical potentials. An example is shown in Fig. 3. In both of these patients the SEPs were recorded within 24 h of the accident and therefore indicate early cessation of brainstem function. Brain death in each was confLrmed within 24 h of head injury. In contrast a P17 component may be present with an absent cortical potential and then subsequently disappear. An example is shown in Fig. 4 where recordings were made on day 4 and day 9 following head injury but no recordings were made on the intervening days. The potential labelled P15 in this illustration is a far-field reflection of activity generated in the medial lemniscus (Momma et al. 1986). The presence of a normal cervical waveform and a P15 recorded at the scalp provides evidence that this patient still had preserved brainstem function on day 4. There is, however, a disruption of
260 DAY 1 RIGHT MEDIAN NERVE
LEFT MEDIAN NERVE
STIMULATED
STIMULATED
C3r- Fpz CORT,CAL SEP
C4'- Fpz
~
~. N13
.~. N13
CERVICAL SEP ~
._•]+
1 pv
4 MS
Fig. 3. Bilateral absence of both the cortical SEP and the PI7 component of the cervical SEP. This was
the first and only recordingand the patient died on the same day.
transmission within the fight thalamocortical radiation or sensory cortex. By day 9 conduction within the brainstem pathways had been lost as well. Brain death was diagnosed within 4 h of this recording.
Group B There were 24 patients in whom the P17 was preserved bilaterally in the last recording. Of these 8 also had an intact N20 recorded unilaterally or bilaterally. Twenty-two of the 24 patients died at least 20 h after the SEP, with survival up to 75 days. The other 2 patients died within 8 h of an SEP showing a preserved P17. In these treatment was discontinued soon after the last SEP was recorded because the head injury was considered unsurvivable. Evidence for continuing cerebral or brainstem function, in the absence of drug effect, was available in all but 2 patients in group B at the time of, or following, the last SEP. Of the 22 patients in whom EEGs were recorded, 14 were at the time of the last SEP and of these 11 had some cerebral activity. The other 3 EEGs were isoclectric hut 2 of these patients made respiratory efforts following removal of ventilator support. The third patient had received large doses of a narcotic during the 2 h prior to the recording. Of the remainder in this group 3 breathed independently following cessation of artificial ventilation, 2 showed pupiUary reactions, 1 had had large doses of phenoharbitone
261
DAY4
DAY9
P15 CORTICAL .~..A ,= ~ ~ . = - ~ . ~ " ~ .
-
SEP
( C4,- Fpz) ~1.N 1 3 CERVICALsEP ~
~
"~ N I ~ , ~
( CII- Fpz) "~P17
.~+1IJV 4 MS
LEFT
MEDIAN
NERVE
STIMULATED
Fig. 4. Deterioration of the P17 component in a patient where the cortical potential had already been lost. On day 4 the only activity recorded at the scalp is a low amplitude positivity ofbralnstem origin (PI5). The cervical waveform is normal with preservation of the P17 component. On day 9 both the PI5 and PI7 components have disappeared. Brain death was diagnosed within 4 h of the last recording.
immediately prior to the SEP and 2 had persistence of an N20 potential on one side. In 2 patients information is not available concerning brainstem function following the last SEP - in one treatment was discontinued and in the other criteria for brain death were full'died 20 h later. One patient did not fit into either group. At the last recording the P17 was absent on one side and poorly defined on the other and an EEG recorded shortly afterwards was isoelectric. Four hours following the SEP recording clinical brain death was declared and it is probable that this patient was developing brainstem death at the time the SEP was being recorded.
DISCUSSION
The evidence that the P17 component of the SEP has a thalamic origin is summarised elsewhere (Shaw 1985). This study is consistent with this hypothesis. All patients who fulfilled the clinical criteria for brain death and had SEPs recorded in close temporal proximity to the diagnosis also had bilateral absence of the P17 components. In these patients brain death was diagnosed either prior to or within a few hours of the SEP recording. However, the P17 potential was always present bilaterally where there
262
was clinical evidence for persisting brainstem function or where there was electrical activity of cerebral origin recorded on the EEG. Those patients in whom the P17 components were preserved either did not fulfil the criteria for brain death and had active treatment removed because it was felt that the brain injury was unsurvivable or else had evidence of continuing brainstem function and did not die until at least 20 h had elapsed following the final SEP recording. A number of previous studies have analysed data on short-latency SEPs in brain death but in none of these was attention paid to the P17 component. When the illustrations in two of the publications are studied a correlation can be seen between the occurrence of brain death and loss of the P 17 potential (Goldie et al. 1981; Belsh and Chokroverty 1987). Similarly, in a study where serial SEP recordings were made at frequent intervals during the process of brain death the gradual deterioration of the P 17 potential can be observed (Buchner et al. 1988). The present study confirms the previous conclusions that the cervical potential may be preserved following brain death though the cortical potential is always lost. However, while this pattern always occurs in brain death, the absence of the N20 potential does not necessarily indicate brainstem death and therefore this pattern cannot be used in the confhanation of clinical brain death. The value of the P17 potential in this situation is that it is more easily recorded than some of the other components of the SEP when brainstem function is being considered and it allows the loss ofbralnstem function to be quantified in an objective manner. While it is theoretically possible that loss of thalamic function, and therefore the P17, may occur with preservation of more caudal brainstem activity, in the context of head injury and bilaterally absent cortical SEPs the disappearance of the P17 most likely indicates bralnstem as well as thalamic destruction. In addition to the loss of P 17, in most patients there was an associated attenuation in the amplitude of the cervical potential (N13). The aetiology of this is uncertain. It might, for example, represent subtraction from the cervical waveform of the amplitude normally contributed by activity generated in the medial lemniscus. Alternatively it may reflect the continuing rostro-caudal loss of neuronal function which occurs in brain death and the consequent deterioration of the spinal generators of the cervical waveform. Most likely it is a combination of both processes. As with other subcortical potentials generated within the somatosensory system the P17 component is robust and is usually unchanged by sedative and analgesic drugs or with temperature fluctuations. In addition the simultaneous recording of the N13 potential indicates that the afferent volley is being conducted to a level close to the cervico-medullary junction. It can be concluded therefore that the shortqatency SEP, and in particular the presence or absence of the P17 component, offers a reliable electrophysiological measure of brainstem function in suspected brain death.
REFERENCES
Anziska, B.J. and R.Q. Cracco (1980) Short latency somatosensoryevoked potentials in brain dead patients. Arch. Neurol., 37: 222-225.
263 Belsh, J.M. and S. Chokroverty (1987) Short latency somatosensory evoked potentials in brain-dead patients. Electroenceph. din. Neurophysiol., 68: 75-78. Buchner, H., A. Ferbert and W. Hacke (1988) Serial recording of median nerve stimulated subcortical somatosensory evoked potentials (SEPs) in developing brain death. Electroenceph. clin. Neurophysiol, 69: 14-23. Cant. B.R., A.L. Hume, J.A. Judson and N.A. Shaw (1986) The assessment of severe head injury by short-latency somatosensory and brainstem auditory evoked potentials. Electroenceph. din. NeurophysioL, 65: 188-195. Cant, B.R. and N.A. Shaw (1986) Central somatosensory conduction time - Method and clinical applications. In: R.Q. Cracco and I. Bodis-Wollner (Eds.), Evoked Potentials, Alan R. Liss, New York, pp. 58-67. Desmedt, J.E. and G. Cheron (1981) Prevertebral (oesophageal) recording of subcortical somatosensory evoked potentials in man: the spinal P13 component and the dual nature of the spinal generators. Electroenceph. clin. NeurophysioL, 52: 257-275. Garcia-Larrea, L., O. Bertrand, F. Artru, J. Pernier and F. Mangni~re (1987) Brainstem monitoring. II. Preterminal BAEP changes observed until brain death in deeply comatose patients. Electroenceph. clin. Neurophysiol., 68: 446-457. Goldie, W.D., K.H. Chiappa, R.R. Young and E.B. Brooks (1981) Brainstem auditory and short-latency somatosensory evoked responses in brain death. Neurology (NY), 31: 248-256. Guidelines for the determination of death (1981) Report of the medical consultants on the diagnosis of death to the President's Commission for the study of ethical problems in medicine and biomedical and behavioral research. J. Am. Med. Assoc., 246: 2184-2186. Hall, J.W., J.R. Mackey-Hargadine and E.E. Kim (1985) Auditory brainstem response in determination of brain death. Arch. OtolaryngoL, 111: 613-620. Hume, A.L. and B.R. Cant (1981) Central somatosensory conduction after head injury. Ann. Neurol., 10: 411-419. Kaga, K., A. Takamori, T. Mizutani, T. Nagai and R. R. Marsh (1985) The auditory pathology of brain death as revealed by auditory evoked potentials, Ann. Neurol., 18: 360-364. Momma, F., H.I. Sabin, L. Symon and N. M. Branston (1986) Clinical evidence supporting a subthalamic origin of the P15 wave of the somatosensory evoked potentials to median nerve stimulation. Electroenceph. din. Neurophysiol., 67: 134-139. Shaw, N.A. (1985) A thalamic component of the cervical evoked potential in man. Neurosci. Lett., 57: 221-225. Starr, A. (1976) Auditory brain-stem responses in brain death. Brain, 99: 543-554. Trojaborg, W. and E.O. Jorgensen (1973) Evoked cortical potentials in patients with "isoeleetric" EEGs. Electroenceph. clin. Neurophysiol., 35: 301-305.
255
JNS 03062
Somatosensory evoked potentials following severe head injury: loss of the thalamic potential with brain death A . M . Chancellor, R.W. Frith and N.A. Shaw Department of Clinical Neurophysiology, Auckland Hospital, Park Road, Auckland (New Zealand) (Received 19 April, 1988) (Revised, received 18 June, 1988) (Accepted 20 June, 1988)
SUMMARY
The thalamic component (P17) of the short-latency somatosensory evoked potential (SEP) was assessed to determine its usefulness in patients with severe head injury. Subjects were a group of patients admitted to the Auckland Hospital Critical Care Unit who subsequently died from head injury. In all instances where brain death was unequivocally established and a SEP recording made in close temporal proximity to the time of brain death the P17 potential was absent. When there was evidence of continuing brainstem activity and particularly where prolonged survival occurred following the last SEP recording the P17 potential remained intact bilaterally. This study shows that the presence or absence of the thalamic component of the short-latency SEP provides a reliable electrophysiological measure of bralnstem function in patients where brain death has been suspected.
Key words: Brain death; Head injury; Somatosensory evoked potential; Thalamic potential
Correspondence to: Dr. R.W. Frith, Neurologist and Clinical Neurophysiologist, Department of Clinical Neurophysiology,Auckland Hospital, Park Road, Auckland l, New Zealand. 0022-510X/88/$03.50 © 1988 Elsevier Science Publishers B.V. (Biomedical Division)
256 INTRODUCTION With the availability of advanced life support systems information is needed on how best to define brain death in the presence of continued cardiopulmonary function. The diagnosis of brain death is usually made by clinical examination although in some centres angiography or an isoelectric EEG in the absence of drug or hypothermic effects provides useful laboratory conf'trmation. The role of other neurophysiological tests, in particular sensory evoked potentials (SEPs), remains uncertain. Short-latency evoked potentials which arise in the primary sensory cortex or more caudal locations have an advantage over the EEG in remaining stable even when the patient has received sedative drugs (Cant and Shaw 1986). A number of studies have now examined the role of brainstem auditory evoked potentials (BAEPs) in the determination of brain death (Starr 1976; Goldie et al. 1981; Hall et al. 1985; Kaga et al. 1985; Garcia-Larrea et al. 1987). The principal waves of the BAEP arise within the 8th cranial nerve and brainstem auditory pathways and therefore these potentials represent a direct measure of brainstem activity. A major problem complicating the use of the BAEP in the diagnosis of brain death is that, for unequivocal interpretation, wave 1 needs to be preserved to indicate that the afferent volley is being propagated within the auditory nerve. Often, however, the entire BAEP complex is abolished. Under these circumstances no certain conclusions can be drawn about the integrity of brainstem structures as complete absence may indicate acute or pre-existing cochlear or auditory nerve damage (Goldie et al. 1981). In addition to the BAEP there have also been reports of short-latency somatosensory evoked potentials (SEPs) in brain death (Trojaborg and Jorgensen 1973; Goldie et al. 1981; Anziska and Cracco 1980; Belsh and Chokroverty 1987; Buchner et al. 1988). Usually, the short-latency SEP consists of the simultaneous recording of the cortical and cervical SEP and there is consensus that while the cortical SEP is invariably lost in brain death, the cervical SEP may be preserved. This pattern of abnormality does not, however, specify at which level within the central somatosensory pathways damage has occurred. Hence the absence of a cortical potential with a preserved cervical potential cannot be taken as indicating loss of brainstem function. SEP components thought to reflect brainstem function have been identified and their disappearance correlated with brain death (Belsh and Chokroverty 1987). The difficulty with such components is that they are not optimally recorded using a cephalic reference and are often ill-defined or not reliably recorded in normal subjects. The principal component of the cortical SEP is a negativity usually labelled N20 which is assumed to be generated within the primary somatosensory cortex (Cant and Shaw 1986). The cervical component appears to be a complex amalgam of activity arising in the cervical roots, dorsal columns, cuneate nuclei and the medial lemnisci. However, the major component of the cervical SEP, the N13 potential, arises within the interneurones of the dorsal horn (Desmedt and Cheron 1981). Following the downward slope of the major negative potential lies a positive trough which, according to its mean latency, has been labelled P17. It has been suggested that this positivity is a far-field reflection of activity generated in or near the thalamus (Shaw 1985). P17
257 probably reflects synaptic and post-synaptic activity from the ventroposterolateral nucleus of the thalamus. P17 is generally robust, stable and is easily recorded. If the potential does arise in the diencephalon it should be absent bilaterally in all patients with brain death. In addition its loss in sequential recordings may be a useful indicator of continuing rostro-caudal deterioration of cerebral function. SEPs recorded from head-injured patients may be divided into three major types (Fig. 1). Type 1 is a normal SEP with intact cortical and cervical wave forms and a normal central conduction time. The central conduction time (CCT) is calculated by subtracting the latency of the cervical potential from that of the cortical potential (Cant and Shaw 1986). Mean CCT is 5.6 msec in adults and a CCT greater than 7.1 msec is abnormal in this laboratory. Type 2 SEP consists of an absent cortical potential but with preservation of both the N13 and P17 components of the cervical potential. It has been shown previously that a type 2 SEP recorded from head-injured patients either unilaterally or bilaterally is associated with an unfavourable outcome (Cant et al. 1986). The type 3 SEP differs from type 2 in that there is loss of the P17 component with preservation of the N13 potential. The purpose of the present study was to investigate the relationship between the loss of the thalamic potential and brain death. We tested the hypothesis that a type 2 SEP was associated with clinical evidence of continuing brainstem function but that a type 3 SEP indicated brain death.
PATIENTS AND METHODS
Subjects were 36 severely head-injured patients who were admitted consecutively to the Department of Critical Care Medicine at Auckland Hospital and who subsequently died of their head injury. Mean survival time was 7 days (median 3 days, range 16 h to 75 days). There were 29 males and 7 females. Median age was 22 years (range 15-73 years). The initial Glasgow Coma Score was less than 8 in all patients. SEPs were recorded on at least 3 of the in'st 4 days foUowing head injury, if the patient survived, and at variable intervals thereafter. Twenty-five patients satisfied clinical criteria for brain death. The remainder had active support withdrawn following a clinical decision that the brain injury was unsurvivable and most died within a short period of time. Three patients remained in a persistent vegetative state for days to weeks prior to death. The diagnosis of brain death was made according to standard criteria based on guidelines of the President's Commission Report (Guidelines 1981). These were: (a) that the patient was not sedated, paralysed or hypothermic, (b) there was no motor activity in the limbs other than that due to spinal reflexes, (c) brainstem reflexes were absent, including, (d) apnoea in the presence of a Paco~ elevated to 60 mm Hg with maintenance of oxygenation. All patients had a postmortem examination which confumed that death was a consequence of head injury. SEPs were recorded using a Tracor-Northern 3 000 evoked response analyser. The median nerves at the wrists were stimulated independently on each side using square wave pulses of 0.15 msec duration presented at a rate of 4.1/sec. Recording
258 electrodes (platinum alloy needles) were inserted over the cervical spine (CII) and over the contralateral sensory cortex (C 4' or C3', 2 cm posterior to C a or C 3, international 10-20 system). A common mid-forehead reference (Fpz) was used. Cervical and cortical SEPs were recorded simultaneously. Analysis time was 40 msec with the first 4 msec following stimulus onset excluded. Sampling interval was 80 #sec. The bandpass of the recording system was 10 Hz to 3 kHz. Each set of SEPs was the average of 500 samples and all SEPs were replicated and superimposed in order to demonstrate reproducibility. Patients were divided into 2 groups according to whether the thalamic component (P17) of the cervical potential was absent or present in the fmal SEP recorded before death. Those in group A had absence of the P17 potential bilaterally in the final SEP recording before death. In group B, the P17 was preserved bilaterally in the final SEP performed prior to death. The 2 groups were compared to determine if loss of the P 17 potential correlated with the patient's clinical status at the time of the last recording.
RESULTS
Figure 1 illustrates the 3 types of short-latency SEPs recorded from patients with severe head injury. Figure 2 illustrates SEPs recorded on 3 consecutive days from one patient. On days 3 and 4 the 3 principal components, N13, P17 and N20, can be identified. However, the cortical potential is reduced in amplitude and there is an
TYPE NORMAL
t
TYPE SEP
LOSS
2
OF
PRESERVATION AND
TYPE N20 OF
LOSS P17
OF
3
N20
AND
PRESERVATION
OF
P17 Nt3
N13
N20
CORTICAL SEP
~
CSVCAL EP '~t' p 1 ? '~r P 1 7 _ ~ + t ~Jv 4 MS
Fig. 1. Examples of three types of short-latency somatosensory evoked potentials (SEPs) recorded from patients following severe head trauma. In type 1 the SEP is normal. The 3 principal components of the waveforms (N13, P17 and N20) are labelled. The central conduction time is 6.5 msec. Central conduction time cannot be calculated in either type 2 or type 3 SEPs because the cortical waveform has been lost. In type 2 SEPs the P17 is preserved while in type 3 it is absent.
259 D AY 3
J CORTICAL SEP
.
~
~
DAY 4
DAY 5
• N20 ~
~ ~
~
..~.~
( C 4' - Fpz )
N I 3 ,~,
( C ii-
N13
Fpz ) P17 _~+1 uv LEFT
MEDIAN
NERVE
STIMULATED
4 MS
Fig. 2. SEPs recorded on 3 consecutive days. On days 3 and 4 both N20 and P17 are preserved although the central conduction time is prolonged. On day 5 neither PIT nor N20 are present. The patient died approxL,nate]y 4 h after the day 5 recording.
abnormal prolongation in CCT to 8 msec. On the following day (day 5) both the cortical and thalamic components had disappeared bilaterally although N13 persisted. The patient died on the same day as the last recording.
Patient data Group A There were 11 patients in this group in whom the P17 was absent bilaterally in the last recording. In 9 of the group brain death was declared prior to, or occurred within 4 h after, the recording of bilaterally absent P17 and cortical potentials. In the other 2 patients a formal brain death assessment was not made until the following day (within 20 h of the recording) but the clinical notes suggest that brain death had occurred on the day of the recording. All patients in group A had isoelectric EEGs recorded at the time of the final SEP. In 2 patients the first and only recording showed a low amplitude N 13 with absent P17 and cortical potentials. An example is shown in Fig. 3. In both of these patients the SEPs were recorded within 24 h of the accident and therefore indicate early cessation of brainstem function. Brain death in each was confLrmed within 24 h of head injury. In contrast a P17 component may be present with an absent cortical potential and then subsequently disappear. An example is shown in Fig. 4 where recordings were made on day 4 and day 9 following head injury but no recordings were made on the intervening days. The potential labelled P15 in this illustration is a far-field reflection of activity generated in the medial lemniscus (Momma et al. 1986). The presence of a normal cervical waveform and a P15 recorded at the scalp provides evidence that this patient still had preserved brainstem function on day 4. There is, however, a disruption of
260 DAY 1 RIGHT MEDIAN NERVE
LEFT MEDIAN NERVE
STIMULATED
STIMULATED
C3r- Fpz CORT,CAL SEP
C4'- Fpz
~
~. N13
.~. N13
CERVICAL SEP ~
._•]+
1 pv
4 MS
Fig. 3. Bilateral absence of both the cortical SEP and the PI7 component of the cervical SEP. This was
the first and only recordingand the patient died on the same day.
transmission within the fight thalamocortical radiation or sensory cortex. By day 9 conduction within the brainstem pathways had been lost as well. Brain death was diagnosed within 4 h of this recording.
Group B There were 24 patients in whom the P17 was preserved bilaterally in the last recording. Of these 8 also had an intact N20 recorded unilaterally or bilaterally. Twenty-two of the 24 patients died at least 20 h after the SEP, with survival up to 75 days. The other 2 patients died within 8 h of an SEP showing a preserved P17. In these treatment was discontinued soon after the last SEP was recorded because the head injury was considered unsurvivable. Evidence for continuing cerebral or brainstem function, in the absence of drug effect, was available in all but 2 patients in group B at the time of, or following, the last SEP. Of the 22 patients in whom EEGs were recorded, 14 were at the time of the last SEP and of these 11 had some cerebral activity. The other 3 EEGs were isoclectric hut 2 of these patients made respiratory efforts following removal of ventilator support. The third patient had received large doses of a narcotic during the 2 h prior to the recording. Of the remainder in this group 3 breathed independently following cessation of artificial ventilation, 2 showed pupiUary reactions, 1 had had large doses of phenoharbitone
261
DAY4
DAY9
P15 CORTICAL .~..A ,= ~ ~ . = - ~ . ~ " ~ .
-
SEP
( C4,- Fpz) ~1.N 1 3 CERVICALsEP ~
~
"~ N I ~ , ~
( CII- Fpz) "~P17
.~+1IJV 4 MS
LEFT
MEDIAN
NERVE
STIMULATED
Fig. 4. Deterioration of the P17 component in a patient where the cortical potential had already been lost. On day 4 the only activity recorded at the scalp is a low amplitude positivity ofbralnstem origin (PI5). The cervical waveform is normal with preservation of the P17 component. On day 9 both the PI5 and PI7 components have disappeared. Brain death was diagnosed within 4 h of the last recording.
immediately prior to the SEP and 2 had persistence of an N20 potential on one side. In 2 patients information is not available concerning brainstem function following the last SEP - in one treatment was discontinued and in the other criteria for brain death were full'died 20 h later. One patient did not fit into either group. At the last recording the P17 was absent on one side and poorly defined on the other and an EEG recorded shortly afterwards was isoelectric. Four hours following the SEP recording clinical brain death was declared and it is probable that this patient was developing brainstem death at the time the SEP was being recorded.
DISCUSSION
The evidence that the P17 component of the SEP has a thalamic origin is summarised elsewhere (Shaw 1985). This study is consistent with this hypothesis. All patients who fulfilled the clinical criteria for brain death and had SEPs recorded in close temporal proximity to the diagnosis also had bilateral absence of the P17 components. In these patients brain death was diagnosed either prior to or within a few hours of the SEP recording. However, the P17 potential was always present bilaterally where there
262
was clinical evidence for persisting brainstem function or where there was electrical activity of cerebral origin recorded on the EEG. Those patients in whom the P17 components were preserved either did not fulfil the criteria for brain death and had active treatment removed because it was felt that the brain injury was unsurvivable or else had evidence of continuing brainstem function and did not die until at least 20 h had elapsed following the final SEP recording. A number of previous studies have analysed data on short-latency SEPs in brain death but in none of these was attention paid to the P17 component. When the illustrations in two of the publications are studied a correlation can be seen between the occurrence of brain death and loss of the P 17 potential (Goldie et al. 1981; Belsh and Chokroverty 1987). Similarly, in a study where serial SEP recordings were made at frequent intervals during the process of brain death the gradual deterioration of the P 17 potential can be observed (Buchner et al. 1988). The present study confirms the previous conclusions that the cervical potential may be preserved following brain death though the cortical potential is always lost. However, while this pattern always occurs in brain death, the absence of the N20 potential does not necessarily indicate brainstem death and therefore this pattern cannot be used in the confhanation of clinical brain death. The value of the P17 potential in this situation is that it is more easily recorded than some of the other components of the SEP when brainstem function is being considered and it allows the loss ofbralnstem function to be quantified in an objective manner. While it is theoretically possible that loss of thalamic function, and therefore the P17, may occur with preservation of more caudal brainstem activity, in the context of head injury and bilaterally absent cortical SEPs the disappearance of the P17 most likely indicates bralnstem as well as thalamic destruction. In addition to the loss of P 17, in most patients there was an associated attenuation in the amplitude of the cervical potential (N13). The aetiology of this is uncertain. It might, for example, represent subtraction from the cervical waveform of the amplitude normally contributed by activity generated in the medial lemniscus. Alternatively it may reflect the continuing rostro-caudal loss of neuronal function which occurs in brain death and the consequent deterioration of the spinal generators of the cervical waveform. Most likely it is a combination of both processes. As with other subcortical potentials generated within the somatosensory system the P17 component is robust and is usually unchanged by sedative and analgesic drugs or with temperature fluctuations. In addition the simultaneous recording of the N13 potential indicates that the afferent volley is being conducted to a level close to the cervico-medullary junction. It can be concluded therefore that the shortqatency SEP, and in particular the presence or absence of the P17 component, offers a reliable electrophysiological measure of brainstem function in suspected brain death.
REFERENCES
Anziska, B.J. and R.Q. Cracco (1980) Short latency somatosensoryevoked potentials in brain dead patients. Arch. Neurol., 37: 222-225.
263 Belsh, J.M. and S. Chokroverty (1987) Short latency somatosensory evoked potentials in brain-dead patients. Electroenceph. din. Neurophysiol., 68: 75-78. Buchner, H., A. Ferbert and W. Hacke (1988) Serial recording of median nerve stimulated subcortical somatosensory evoked potentials (SEPs) in developing brain death. Electroenceph. clin. Neurophysiol, 69: 14-23. Cant. B.R., A.L. Hume, J.A. Judson and N.A. Shaw (1986) The assessment of severe head injury by short-latency somatosensory and brainstem auditory evoked potentials. Electroenceph. din. NeurophysioL, 65: 188-195. Cant, B.R. and N.A. Shaw (1986) Central somatosensory conduction time - Method and clinical applications. In: R.Q. Cracco and I. Bodis-Wollner (Eds.), Evoked Potentials, Alan R. Liss, New York, pp. 58-67. Desmedt, J.E. and G. Cheron (1981) Prevertebral (oesophageal) recording of subcortical somatosensory evoked potentials in man: the spinal P13 component and the dual nature of the spinal generators. Electroenceph. clin. NeurophysioL, 52: 257-275. Garcia-Larrea, L., O. Bertrand, F. Artru, J. Pernier and F. Mangni~re (1987) Brainstem monitoring. II. Preterminal BAEP changes observed until brain death in deeply comatose patients. Electroenceph. clin. Neurophysiol., 68: 446-457. Goldie, W.D., K.H. Chiappa, R.R. Young and E.B. Brooks (1981) Brainstem auditory and short-latency somatosensory evoked responses in brain death. Neurology (NY), 31: 248-256. Guidelines for the determination of death (1981) Report of the medical consultants on the diagnosis of death to the President's Commission for the study of ethical problems in medicine and biomedical and behavioral research. J. Am. Med. Assoc., 246: 2184-2186. Hall, J.W., J.R. Mackey-Hargadine and E.E. Kim (1985) Auditory brainstem response in determination of brain death. Arch. OtolaryngoL, 111: 613-620. Hume, A.L. and B.R. Cant (1981) Central somatosensory conduction after head injury. Ann. Neurol., 10: 411-419. Kaga, K., A. Takamori, T. Mizutani, T. Nagai and R. R. Marsh (1985) The auditory pathology of brain death as revealed by auditory evoked potentials, Ann. Neurol., 18: 360-364. Momma, F., H.I. Sabin, L. Symon and N. M. Branston (1986) Clinical evidence supporting a subthalamic origin of the P15 wave of the somatosensory evoked potentials to median nerve stimulation. Electroenceph. din. Neurophysiol., 67: 134-139. Shaw, N.A. (1985) A thalamic component of the cervical evoked potential in man. Neurosci. Lett., 57: 221-225. Starr, A. (1976) Auditory brain-stem responses in brain death. Brain, 99: 543-554. Trojaborg, W. and E.O. Jorgensen (1973) Evoked cortical potentials in patients with "isoeleetric" EEGs. Electroenceph. clin. Neurophysiol., 35: 301-305.