British Journal of Anaesthesia 85 (3): 471±3 (2000)

CASE REPORTS
Repetitive synchronized cyclical oscillations of multisystem
parameters subsequent to high-dose thiopental therapy for status
epilepticus secondary to herpes encephalitis
I. R. Ghosh1 2 *, R. M. Langford2, K. Nieminen3, A. Kari3 and J. Takala3
1

Department of Clinical Neurophysiology and 2Anaesthetic Laboratory, St Bartholomew's Hospital,
London, UK. 3Critical Care Research Program, Department of Anaesthesiology and Intensive Care,
Kuopio University Hospital, Kuopio, Finland
*Corresponding author: Anaesthetic Laboratory, St Bartholomew's Hospital, 38 Little Britain, London EC1A 7BE, UK
We report a case of status epilepticus secondary to herpes encephalitis, treated with thiopental
infusion and mechanical ventilation. The computerized storage and analysis of physiological data
led to the detection of repetitive synchronized cyclical oscillations of arterial pressure, heart
rate, EEG parameters, peripheral temperature and core temperature. Arterial pressure oscillations have been described in patients who are severely systemically unwell; cardiovascular and
brain electrical activity may also oscillate in the presence of raised intracranial pressure. In contrast, this patient had no features of severe systemic illness or of raised intracranial pressure.
Our hypothesis is that high-dose thiopental may have been a cause of our ®ndings by producing
autonomic dysfunction.
Br J Anaesth 2000; 85: 471±3
Keywords: intensive care; complications, epilepsy; anaesthetics i.v., thiopentone; monitoring;
monitoring, electrocardiography; monitoring, electroencephalography
Accepted for publication: April 14, 2000

Case report
An 18-yr-old male with no history of epilepsy was treated in
the emergency unit for status epilepticus. There were no
focal neurological signs or evidence of raised intracranial
pressure. Seizure activity persisted despite cumulative total
doses of phenytoin 1500 mg and diazepam 60 mg. A loading
dose of thiopental 1000 mg was therefore given followed by
tracheal intubation and mechanical ventilation. Positive
herpes zoster antibody serology and cerebrospinal ¯uid
polymerase chain reaction studies later con®rmed the
diagnosis of viral encephalitis.
On admission to the intensive care unit (ICU), intermittent positive-pressure ventilation was continued, together
with a thiopental infusion titrated up to 500 mg h±1 under the
guidance of a continuously running paper electroencephalogram (EEG), which con®rmed the presence of burst
suppression and the absence of seizure activity. Standard
monitoring included continuous electrocardiography
(ECG), invasive arterial pressure, capnography, respiratory
¯ow/volume measurements, hourly urine output measurements and repeated arterial blood gas analysis. The

cardiorespiratory, haematological, hepatic, renal, biochemical and blood gas pro®les were within normal limits. In
particular, no signs of poor tissue oxygenation or pH
imbalance were observed. Cranial CT and MRI scans
showed no abnormalities. Additional therapy included
acyclovir 750 mg 8-hourly and a crystalloid infusion at
100±250 ml h±1. No sympathomimetic therapy was
required. No subsequent seizures were detected clinically
or electroencephalographically, and the thiopental infusion
was stopped on day 2, by which time the patient had
received 7975 mg over 20 h. Regular intravenous phenytoin
therapy (250 mg 12-hourly) was continued.
On day 3, with the relatives' assent and research ethics
committee approval, specialized monitoring was set up as
part of the EU Biomed 1 IMPROVE intensive care project1
in order to establish an annotated data library. This involved
acquiring and storing continuous physiological variables
(including arterial pressure, ECG and capnographs at
sampling rates between 25 and 100 Hz) and intermittent
annotations (exact timing of drug doses, insertion of lines
and any other interventions). This period of specialized

Ó The Board of Management and Trustees of the British Journal of Anaesthesia 2000

Ghosh et al.

Fig 1 Twenty-four-hour display of polygraphic physiological measurements. The variables are shown on separate graphs for clarity. (A) Root mean
square EEG amplitude, a frequently used processing technique in which digitized values for EEG amplitude are individually squared within a sampled
period; subsequently the mean is extracted and the square root of the mean is taken. (B) EEG median power frequency, the frequency at which half the
spectral power is above and half is below within a sampled period. (C) Mean arterial pressure. (D) Heart rate. (E) Core temperature. (F) Urine output.

monitoring lasted 24 h, during which time continuous
digitized EEG was also recorded.2 No notable clinical
events occurred in this 24 h period, and biochemical and
blood gas pro®les remained within normal limits. As on
admission, there were no signs of raised intracranial
pressure.
Subsequently, the patient made an uncomplicated recovery; his level of consciousness improved and the tracheal
tube was removed on day 6. At this point his Glasgow Coma
Scale score was 15 and gross intellectual functions were
normal. He left the ICU on day 7 and went home on day 14.
Retrospective trend analysis revealed marked cyclical
oscillations of the multisystem physiological variables
over 24 h (Fig. 1). Assessments were made of raw data
(peripheral and core temperature, mean arterial pressure,
heart rate) and frequently used processed EEG parameters. The latter consisted of root mean squared
amplitude (each EEG amplitude digital value within a

sampling period is squared, a mean is subsequently
taken and the square root of this is taken) and median
power frequency (frequency at which half the spectral
power is above and half is below). Peaks of relative
hypertension occurred synchronously with peaks in EEG
amplitude. Oscillations in heart rate and EEG frequency
both showed a close inverse relationship to arterial
pressure and EEG amplitude. Peripheral temperature and
urine output also showed oscillations, though these did
not show exact synchrony with the other parameters.

Discussion
The variability of physiological parameters is an important
indicator used in assessment and prediction in intensive care
patients. Several types of variability have been observed:
the simplest example is the ECG R±R interval, which is a
good measure of (i) reactivity to stress during critical illness

472

Oscillations of multisystem parameters after thiopental

and (ii) autonomic neuropathy. Similarly, EEG reactivity to
stimuli carries the same implications; the presence of
reactivity is reassuring, and the lack of reactivity (in the
absence of drug-induced or hypothermic coma) suggests a
poor prognosis.3
More orderly variations exist in both cardiac and
neurological signals. First-, second-, third- and, more
recently, fourth-order blood-pressure waves have been
described and summarized elegantly by Seiver and colleagues.4 First- and second-order waves are descriptions of
blood-pressure variations with each systolic beat and
respiratory cycle, respectively. Third-order waves, with a
cycle of 10±160 s, are a consequence of disordered
autonomic feedback. Fourth-order waves (cycles of minutes
to hours, as seen in this case) have been thought to be due to
poor oxygen delivery to the tissues and are associated with a
poor prognosis.4 5 Electroencephalographers have described
cyclical changes in association with sleep patterns;6 the
presence or absence of these features has, in turn, relevance
to prognostication in coma.7 8 Repetitive waves with cycling
of minutes to hours are also seen in the EEG monitoring of
patients with brain insult as a consequence of raised
intracranial pressure.9
In this case, we considered three possible causes of the
multiparameter oscillations. Classical fourth-order changes
seemed unlikely, as severe systemic dysfunction was not
evident. Covertly raised intracranial pressure may have been
an alternative cause, though there was never any clinical or
radiological evidence of it, or of persisting brain injury.
Lastly, we could not exclude high-dose thiopental infusion
as an important cause. Plasma levels were not measured, as
these ®ndings became apparent only on trend analysis after
the cessation of monitoring. Our hypothesis is that highdose thiopental may cause autonomic dysfunction, leading
to loss of `randomicity'; this would be similar to the
hypothesis of fourth-order oscillations caused by poor tissue
oxygen delivery in ill patients.
The interesting phenomenon of synchronized cyclical
oscillations of physiological parameters should not automatically instil concern, and may prove to be a recognized
®nding when high-dose thiopental infusion is used or there

is covert brain dysfunction. This may, in time, be clari®ed
with the increasing sophistication of polygraphic trend
analysis in intensive care, without which our ®ndings would
not have been apparent. To validate these ®ndings,
prospective studies of high-dose thiopental infusions
(including the simultaneous measurement of plasma levels)
and intracranial pressure monitoring should be encouraged.

Acknowledgements
We acknowledge the help of the Kuopio ICU staff and the loan of
equipment from Aalborg University and St Bartholomew's Hospital,
London (Anaesthetics Laboratory and Clinical Neurophysiology
Department).

References

473

1 Carson E, Saranummi N. Improving control of patient status in
critical care: the IMPROVE project. Comp Methods Prog Biomed
1996; 51: 1±130
2 Thomsen CE, Gade J, Nieminen K, Langford RM, Ghosh IR,
Jensen K, et al. Collecting EEG Signals in the IMPROVE Data
Library. IEEE Eng Med Biol 1997; 16: 33±40
3 Synek VM. Prognostically important EEG coma patterns in diffuse
anoxic and traumatic encephalopathies in adults. J Clin
Neurophysiol 1988; 5: 161±74
4 Seiver A, Daane S, Kim R. Regular low frequency cardiac output
oscillations observed in critically ill surgical patients. Complexity
1997; 2: 51±5
5 Carroll G. Fourth-order blood pressure waves. J Am Med Assoc
1990; 236: 856
6 Loomis AL, Harvey EN, Hobart G. Potential rhythms of the
cerebral cortex during sleep. Science 1938; 81: 589±97
7 Chatrian GE, White LE, Daly D. Electroencephalographic
patterns resembling those of sleep in certain comatose states
after injuries to the head. Electroencephalogr Clin Neurophysiol
1963; 15: 272±9
8 Evans BM, Bartlett JR. Prediction of outcome in severe head
injury based on recognition of sleep related activity in the
polygraphic electroencephalogram. J Neurol Neurosurg Psychiatry
1995; 59: 17±25
9 Munari C, Calbucci F. Correlations between intracranial
pressure and EEG during coma and sleep. Electroencephalogr
Clin Neurophysiol 1981; 51: 170±6

British Journal of Anaesthesia 85 (3): 474±6 (2000)

Loss of consciousness following spinal anaesthesia for Caesarean
section
Y. K. Chan*, R. Gopinathan and R. Rajendram
Department of Anaesthesiology, Faculty of Medicine, University of Malaya Medical Centre, 50603 Kuala
Lumpur, Malaysia
*Corresponding author
A healthy parturient under spinal anaesthesia for Caesarean section lost consciousness for an
hour, 20 min after the intrathecal injection of 2 ml of 0.5% heavy bupivacaine. The patient was
haemodynamically stable before losing consciousness. The differential diagnosis is discussed.
Br J Anaesth 2000; 85: 474±6
Keywords: anaesthesia, depth; anaesthetic techniques, subarachnoid; anaesthesia, obstetric
Accepted for publication: April 14, 2000

Spinal anaesthesia is increasingly used for Caesarean
section1 as it is remarkably safe and effective.2 Its safety3
lies in the fact that parturients are conscious throughout
anaesthesia and are not exposed to the hazards of intubation.
Patients can lose consciousness in the presence of a total
spinal block,4 usually characterized by severe hypotension.5
However, loss of consciousness with this technique in a
haemodynamically stable patient is unusual. We report a
parturient who lost consciousness 20 min after a successful
spinal anaesthetic block.

Case report
The parturient, a 25 yr old, 160 cm, 55 kg normotensive
primigravida with mild polyhydramnios, required an elective Caesarean section for delivery of a baby with bilateral
pleural effusions of unknown cause. She had no other
medical condition of note; in particular, she had no
psychiatric history. Plasma electrolytes were normal before
surgery. Sodium citrate and oral ranitidine were given in the
ward before delivery; no other drugs were administered and
the patient was not on any medication. Immediately before
spinal blockade, Unasyn 1.5 g (subactam 500 mg and
ampicillin 1000 mg) diluted in sterile water was administered as an intravenous infusion. For spinal block, a 26 G
Atraucan (B Braun) spinal needle was used at L3/4 in the
midline with the patient in the sitting position.
Cerebrospinal ¯uid ¯owed from the needle after a single
attempt and 2 ml of 0.5% heavy bupivacaine was injected
after initial aspiration. The patient was then placed
horizontally with a left lateral tilt and a single pillow to
support the head. A T4 level of block was obtained by 5 min
and the Bromage score before surgery was zero. Before
delivery of the baby, the parturient's arterial pressure was

measured every 5 min with a Dinamap. This averaged
around 110/60 mm Hg with one transient episode of
hypotension (90/50 mm Hg) about 5 min after the spinal
injection, which improved with the administration of
ephedrine 5 mg. The heart rate averaged 80±100 beats
min±1. Three litres of oxygen were administered by nasal
prongs; oxygen saturation was 99±100%. Hartmann's
solution (500 ml) was given as a preload and a second
500 ml was commenced before the baby was delivered.
During this period the patient was alert and communicating
with the anaesthetist. No other drugs were administered for
15 min before delivery of the baby.
At the time of delivery, 20 min after administration of the
intrathecal drug, the patient suddenly stopped communicating, as if going off to sleep. She was not responding to
verbal commands or to deep pain. There was no frothing,
and no uprolling of the eyeballs. The incident was not
preceded by nausea and vomiting. There were no complaints of chest pain, inability to breathe or weakness of the
upper limbs immediately before loss of consciousness.
There was a transient period of apnoea with a dramatic drop
in oxygen saturation to around 75%. The patient's lungs
were ventilated with a face mask for 1 min on the circle
system after which spontaneous ventilation resumed. No
bronchospasm or urticaria was noted. The anaesthetist tried
to intubate the patient at this stage but was unsuccessful as
the patient had a very strong gag re¯ex. A decision was
made not to give additional drugs to sedate or paralyse the
patient as it was noted that she was able to protect her
airway. Both of her pupils were about 3 mm wide and
reactive to light. Throughout this period of loss of
consciousness, arterial pressure averaged around 100/55
mm Hg and the pulse rate was about 80 min±1. Ten minutes
into the incident the patient was responding to deep pain

Ó The Board of Management and Trustees of the British Journal of Anaesthesia 2000

Spinal anaesthesia for Caesarean section

with non-purposeful movements of the upper limbs.
Gradually the movements on deep pain became more
purposeful but there was still no verbal response. One hour
after the onset of the incident, the patient opened her eyes to
commands and had slurred speech when she attempted to
talk. Muscle power in the upper limbs at this stage was grade
4 and her sensory level was at T2 bilaterally. The blood
sugar concentration was 3.9 mmol litre±1. One and a half
hours after the onset of the loss of consciousness, the patient
was fully alert and her upper limb muscle power was grade 5
with a sensory level at T5. When questioned about her loss
of consciousness, she said that she had felt extremely
sleepy, had gone to sleep and was unable to recall the event.
Four hours after loss of consciousness, the spinal block had
totally worn off, with grade 5 muscle power in all four
limbs. The patient was discharged back to the ward where
the anaesthetist visited her for the next few days. The baby
died because of severe heart failure, with a single atrium,
single ventricle and pulmonary atresia, 2 h after delivery.
Twenty-four hours after spinal injection, the patient
complained of giddiness and a postural frontal headache
which gradually worsened. CT scan results were normal and
an EEG showed no abnormal foci. An epidural blood patch
relieved the headache immediately. The patient was
discharged 1 week after the incident with no further
complaints.

Discussion
A non-sedated parturient given an intrathecal local anaesthetic injection should never lose consciousness. In this
particular parturient, there was no history of epilepsy or
diabetes to cause loss of consciousness. A blood sugar
concentration of 3.9 mmol litre±1 is at the low end of the
normal range but was not low enough to account for the loss
of consciousness. The normal EEG indicated that the event
could not be explained by an epileptic seizure. A CT scan
done when the patient developed a headache 24 h after the
incident was normal. This indicated that there was no spaceoccupying lesion intracranially to account for the loss of
consciousness during the Caesarean section. As magnetic
resonance imaging (MRI) is superior to CT scanning for
evaluating most suspected central nervous system pathology,6 we should, in retrospect, have subjected the patient
to MRI instead. Only 2 ml of 0.5% bupivacine was given
intrathecally and the patient was subsequently placed
horizontally, so the signs could not be attributed to a total
spinal block. The patient did not develop signi®cant
hypotension, respiratory depression or pupillary dilation to
warrant a diagnosis of total spinal block.4 5 The anaesthetist
was certain that the loss of consciousness could not be
attributed to any intravenous sedative drugs being given
inadvertently, as no drugs were administered for 15 min
before delivery of the baby. It was unlikely that the patient
had an amniotic ¯uid embolism,7 although the event
occurred at the time of delivery of the baby. She was stable

cardiovascularly and at no time was there any coagulopathy.
The event was also unlikely to be the result of a pulmonary
embolism,8 as the patient had a normal cardiac rhythm.
There was no cardiovascular compromise at the time of the
event and no residual neurological de®cit after regaining
consciousness. However, echocardiography was not attempted. Air embolism occurring at the time of delivery can
also be responsible for the loss of consciousness. For loss of
consciousness to have occurred, air must have reached the
left side of the heart through a patent foramen ovale. This
was not ruled out by echocardiography, but air embolism is
unlikely as there were no complaints of dyspnoea or chest
pain,9 nor was there residual neurological de®cit beyond the
period of loss of consciousness. The patient described by
Davis, Glover and Maycock,10 with cerebral arterial air
embolism, not only had a prolonged recovery from general
anaesthesia but had a neurological de®cit for several days
despite hyperbaric oxygen therapy.
Skowronski and Rigg11 and Philip and Walter12 described
loss of consciousness in patients following the administration of epidural anaesthesia for labour. In both these
parturients, the loss of consciousness had occurred, as in
our patient, in the presence of remarkable haemodynamic
stability. These were described as total spinal anaesthesia.
These complications were probably the result of subdural
spread. The signs that alerted the anaesthetists to the cranial
extension in these case reports were dif®culty in breathing,11
arm weakness12 and dysarthria.12 These signs developed
slowly with time and the anaesthetists involved were able to
detect them. In our case, the loss of consciousness after 20
min of relatively uneventful spinal blockade was the ®rst
sign of a possible cranial extension. The loss of consciousness precluded detailed sensory and motor testing in the
head, but the gradual and progressive return of function
indicated that it could have been compatible with anaesthetic blockade that receded with time. Although we do not
have radiological evidence to prove that this incident was
caused by subdural blockade, we believe that the clinical
events that occurred in our patient are in keeping with this
suggestion.5 13 These include relatively stable arterial pressure, slow onset of symptoms after 20 min and complete
recovery after almost 2 h. The absence of sympathetic
blockade14 15 is consistent with subdural blockade.
The parturient could have lost consciousness as a
psychogenic response to distress felt at the time of the
Caesarean section, knowing that her newborn was critically
ill and not expected to survive the neonatal period. Hysteria
cannot be excluded from the differential diagnosis since
measurements taken during the incident were normal except
for a transient episode of hypoxia and the 1 h loss of
consciousness. Normal antenatal and postnatal mental
health makes this diagnosis improbable.
In summary, we have described loss of consciousness in a
parturient who had a successful spinal block. The loss of
consciousness occurred in the presence of remarkable
haemodynamic stability. Despite investigations, we were

475

Duncan et al.

unable to determine the cause. We can only speculate that a
subdural block complicating the spinal block may have
accounted for this most unusual event.

References
1 Brown GW, Russell IF. A survey of anaesthesia for Caesarean
section. Int J Obstet Anesth 1995; 4: 214±8
2 Moore DC. Spinal (subarachnoid) block; a review of 11574 cases.
J Am Med Assoc 1966; 195: 907±12
3 Report on con®dential enquiries into maternal deaths in the
United Kingdom 1988±1990. London, HMSO, 1994.
4 Gillies IDS, Morgan M. Accidental total spinal analgesia with
bupivacaine. Anaesthesia 1973; 28: 441±5
5 Collier C. Total spinal or massive subdural block? Anaesth Intens
Care 1982; 10: 92±3
6 Quint DJ. Indications for emergent MRI of the central nervous
system. J Am Med Assoc 2000; 283: 853±5
7 Clark SL. New concepts of amniotic ¯uid embolism: a review.
Obstet Gynecol Surv 1990; 45: 360±8
8 Dehring DJ, Arens JF. Pulmonary thromboembolism: Disease

recognition and patient management. Anesthesiology 1990; 73:
146±64
9 Malinow AM, Naulty JS, Hunt CO, Datta S, Ostheimer G.
Precordial ultrasonic monitoring during Cesarean delivery.
Anesthesiology 1987; 66: 816±9
10 Davis FM, Glover PW, Maycock E. Hyperbaric oxygen for
cerebral arterial air embolism occuring Caesarean section.
Anaesth Intens Care 1990; 18: 403±5
11 Skowronski GA, Rigg JRA. Total spinal block complicating
epidural analgesia in labour. Anaesth Intens Care 1981; 9: 274±6
12 Philip JH, Walter UB. Total spinal anesthesia late in the course of
obstetric bupivacaine epidural block. Anesthesiology 1976; 44:
340±1
13 Reynolds F, Speedy HM. The subdural space: the third place to
go astray. Anaesthesia 1990; 45: 120±3
14 Pearson RMG. A rare complication of extradural analgesia.
Anaesthesia 1984; 39: 460±3
15 Stevens RA, Stanton-Hicks MD. Subdural injection of local
anesthetic: a complication of epidural anesthesia. Anesthesiology
1985; 63: 325±6

British Journal of Anaesthesia 85 (3): 476±8 (2000)

Traumatic bilateral internal carotid artery dissection following
airbag deployment in a patient with ®bromuscular dysplasia
M. A. Duncan*, N. Dowd, D. Rawluk and A. J. Cunningham
Department of Anaesthesia, Intensive Care Medicine and Neurosurgery, Beaumont Hospital, Beaumont
Road, Dublin 9, Ireland
*Corresponding author
This case describes a 39-yr-old male, presenting with left hemiplegia after a road traf®c accident
involving frontal deceleration and airbag deployment. Brain computerized tomography (CT)
scan revealed a right parietal lobe infarct. Contrast angiography demonstrated bilateral internal
carotid artery dissection and ®bromuscular dysplasia. The patient was treated with systemic
heparinization. Neurological improvement, evidenced by full return of touch sensation, proprioception and nociception began 10 days after the injury. To our knowledge, this is the ®rst
case report of carotid artery dissection associated with airbag deployment. Forced neck extension in such settings may result in carotid artery dissection because of shear force injury at the
junction of the extracranial and intrapetrous segments of the vessel. Clinicians should consider
carotid artery injury when deterioration in neurological status occurs after airbag deployment.
We propose that the risk of carotid artery dissection was increased by the presence of ®bromuscular dysplasia.
Br J Anaesth 2000; 85: 476±8
Keywords: complications, accidents; complications, trauma; complications, ®bromuscular
dysplasia; arteries, carotid
Accepted for publication: April 14, 2000

Ó The Board of Management and Trustees of the British Journal of Anaesthesia 2000

Carotid injury and airbag deployment

Airbags were developed to decrease the incidence of death
and severe injury following road traf®c accidents.1 Airbags
are reported to cause injuries in 43% of deployments. The
majority of these are minor (abrasions and contusions), but
severe injury such as upper limb fracture occurs in 4% of
cases.1 This case report describes bilateral internal carotid
artery dissection, associated with airbag deployment, in a
patient with ®bromuscular dysplasia.

Case report
A 39-yr-old male driver, height 165 cm, was involved in an
acute deceleration, frontal-impact collision. He was restrained by both a seatbelt and an airbag. He was transferred
by road ambulance to the Accident & Emergency unit at our
institution. On initial examination the patient was haemodynamically stable. His Glasgow Coma Score (GCS) was
10/15 at the time of initial assessment but improved to 13/15
thereafter. Neurological examination revealed a left homonymous hemianopia, left hemiplegia and an upper motor
neurone left 7th nerve palsy. Other injuries sustained
included metatarsal fractures and anterior neck abrasions.
Standard haematological and biochemical investigations
and chest and cervical spine radiographs were normal. A
brain computerized tomography (CT) scan showed a right
parietal lobe infarct.
The patient was transferred to the neurosurgical intensive
care unit for invasive arterial pressure monitoring and
neurological observations. Supplemental oxygen was
administered by facemask. Aspirin 300 mg once daily by
mouth was prescribed for its antiplatelet effect. A second
brain CT on hospital day 2 demonstrated haemorrhagic
transformation of the right parietal lobe infarct. Duplex scan
and magnetic resonance angiogram of the carotid arteries
was inconclusive. Contrast angiography revealed bilateral
internal carotid artery dissection, ®bromuscular dysplasia of
the internal carotid arteries and thrombus in the right
internal carotid artery (Fig. 1). Heparin was commenced
with a target activated partial thromboplastin time of 60 s.
Warfarin was commenced with a target prothrombin time of
26 s.
Recovery of neurological function began with the return
of normal GCS on hospital day 3. By day 10, touch
sensation, proprioception and nociception in the left upper
and lower limbs were evident. Left lower limb power had
improved to grade 3/5 on day 24. However, the left faciobrachial palsy persisted. The patient was transferred to a
rehabilitation hospital on day 26.

Discussion
Blunt carotid injury presents typically with neck abrasions
and neurological sequelae, including hemiplegia, low GCS
and a co-existing Horner's syndrome when the sympathetic
chain is stretched across the transverse process of the ®rst
cervical vertebrae.2 3 Road traf®c accidents are the most

Fig 1 Right carotid angiogram showing intimal dissection (1) at the
junction of the extracranial and intrapetrous segments of the internal
carotid artery, and the `string of beads' appearance (2) that is
characteristic of ®bromuscular dysplasia.

frequent cause of blunt carotid injury.3 The mechanism of
injury to the internal carotid artery is thought to be a shear
force applied to the vessel during forced neck extension,
causing intimal dissection at the junction of the immobile
intracranial and relatively mobile extracranial portions of
the internal carotid artery.2
Fibromuscular dysplasia is an arterial developmental
abnormality of unknown aetiology affecting medium-sized
arteries.4 5 The incidence of the disorder based on a series of
20 244 post-mortems is two per 1000 population.4 There are
four known subtypes, each with its own unique histological
appearance (intimal ®broplasia, medial ®broplasia, medial
hyperplasia and perimedial dysplasia).5 Affected vessels are
elongated, kinked and punctuated with stenoses and
microaneurysms.5 These abnormalities produce unique
angiographic features which radiologists call the `string of
beads' appearance (Fig. 1). The presence of ®bromuscular
dysplasia may compromise the structural integrity of the
affected vessels, including the carotid artery, and therefore
predispose to vessel dissection after traumatic injury.
Patients with ®bromuscular dysplasia are asymptomatic
until they sustain a vessel dissection. They then present with

477

Duncan et al.

ischaemia of the organs supplied by the affected vessel, for
example cerebrovascular accident. Fibromuscular dysplasia
is found in 15% of cases of spontaneous carotid dissection.2 6 A report of bilateral internal carotid and vertebral
artery dissection in a patient with ®bromuscular dysplasia
after a road traf®c accident has been described.7
Arteriography is the gold standard diagnostic test for both
®bromuscular dysplasia and blunt carotid injury.5 8
The management of carotid artery dissection is controversial. Heparinization was associated with an improved
outcome in the largest series of blunt carotid injuries.3
However, the safety and ef®cacy of anticoagulation in the
management of stroke following carotid injury has not been
demonstrated.9 In our patient, the apparent bene®t of
heparinization had to be balanced against the risks of a
worsening stroke.
Airbag usage is increasing worldwide and is now
mandatory in new cars in some countries.1 Airbags function
by preventing traumatic impact between the vehicle occupant and the vehicle interior. During a road traf®c accident,
sensors in the vehicle body detect deceleration and trigger
airbag deployment.1 To protect against injury, airbags must
be fully expanded before the vehicle occupant is propelled
forward by the deceleration of the accident. An airbag
expansion speed of 200 mph is necessary. This rapid
forceful expansion causes airbag-associated injury in 43%
of deployments. In addition, fatal rupture of the ventral
ligaments of the cervical spine and brain stem has been
described secondary to forced neck extension.10
An increase in the incidence of airbag-associated injury is
predicted.1 Drivers of short stature, as in this case, are at
increased risk of these injuries.8 9 Their short stature makes
it necessary to move the vehicle seat forward in order to
reach the foot pedals, thereby bringing the upper body into
the path of the expanding airbag. In the USA it is now
recommended that airbags should be decommissioned if
drivers cannot keep a wheel±chest distance of 10 inches.1
The evidence for airbag-mediated blunt carotid injury by
means of forced neck extension is compelling in this case. A
less likely explanation is that the patient initially sustained
spontaneous carotid artery dissection because of ®bromuscular dysplasia and then crashed. However, spontaneous
dissections of the carotid artery typically present initially

with minor clinical manifestations (e.g. headache),10 and
only develop neurological signs after several days.11 12
In summary, this is the ®rst case report of bilateral
internal carotid artery dissection, in a patient with
®bromuscular dysplasia, associated with airbag deployment.
The report suggests that clinicians should consider the
possibility of carotid injury when neurological deterioration
occurs after a road traf®c accident with airbag deployment.
Associated conditions, such as ®bromuscular dysplasia,
may increase the risk of carotid artery injury in this setting.
Early carotid angiography is warranted to make the
diagnosis.

References

478

1 Mikhail JN, Huelke DF. Airbags: an update. J Emerg Nursing 1997;
23: 439±45
2 Hart RG, Easton JD. Dissections of cervical and cerebral arteries.
Neurologic Clin 1983; 1: 155±82
3 Fabian TC, Patton JH, Croce MA, Minard G, Kudsk KA, Pritchard
EF. Blunt carotid injury. Ann Surg 1996; 223: 513±25
4 Schievink WI, Bjornsson J. Fibromuscular dysplasia of the internal
carotid artery: a clinicopathological study. Clin Neuropathol 1996;
15: 2±6
5 Porter JM, Taylor LM, Harris EJ Jr. Nonatherosclerotic vascular
disease In: Moore WS, ed. Vascular Surgery: A Comprehensive
Review, 4th edition. Philadelphia: W.B. Saunders, 1993: 108±46
6 Mokri B, Sundt TM Jr, Houser OW, Piepgras DG. Spontaneous
dissection of the cervical internal carotid artery. Ann Neurol
1986; 19: 126±38
7 Eachempati SR, Sebastian MW, Reed RL. Posttraumatic bilateral
carotid artery and right vertebral artery dissections in a patient
with ®bromuscular dysplasia: case report and review of the
literature. J Trauma 1998; 44: 406±9
8 Perry MO. Injuries of the brachiocephalic vessels. In: Rutherford
RB, ed. Vascular Surgery, 4th edition. Philadelphia: W.B. Saunders,
1995: 705±13
9 Hacke W, Stingele R, Steiner T, Schuchardt V, Scwab S. Critical
care of acute ischaemic stroke. Intensive Care Med 1995; 21: 856±
62
10 Maxeiner H, Hahn H. Airbag induced lethal cervical trauma. J
Trauma 1997; 42: 1148±51
11 Perez J, Palmatier T. Airbag-related fatality in a short, forward
positioned driver. Ann Emerg Med 1996; 28: 722±4
12 Biousse V, Anglejan-Chatillon J, Touboul P, Amarenco P, Bousser
M. Time course of symptoms in extracranial carotid artery
dissections. Stroke 1995; 26: 235±9

British Journal of Anaesthesia 85 (3): 479±81 (2000)

Anaesthetic management of a pregnant patient in a persistent
vegetative state
B. T. Ayorinde1, I. Scudamore2 and D. J. Buggy1,*
1

University Department of Anaesthesia and 2Department of Obstetrics and Gynaecology,
University Hospitals of Leicester, Leicester General Hospital, Gwendolen Road, Leicester LE5 4PW, UK
*Corresponding author
Pregnancy in a patient in a persistent vegetative state presents challenging therapeutic questions
about the level of supportive management required, the assessment of fetal well-being, the
timing and mode of delivery and the anaesthetic management of labour and delivery. We
report the case of a 29-yr-old woman who had a favourable fetal outcome despite suffering
hypoxic brain damage after a suicide attempt by a drug overdose. She was managed until the
onset of labour on an intensive care unit and had a spontaneous vaginal delivery assisted by
epidural anaesthesia.
Br J Anaesth 2000: 85; 479±81
Keywords: brain, injury; persistent vegetative state; anaesthesia, epidural; pregnancy
Accepted for publication: April 14, 2000

The term persistent vegetative state (PVS) describes a state
of continuing `wakefulness without awareness' that is
usually due to a severe brain insult from a variety of
possible causes.1 PVS involves damage to the central
nervous system, other systems being essentially normal. A
pregnancy may be viable, creating an unusual scenario
requiring dif®cult clinical and ethical decisions. In this
report, we discuss the anaesthetic and intensive care
management of a 29-yr-old patient who suffered hypoxic
brain injury after a failed suicide attempt at 12 weeks
gestation.

Case report
A 29-yr-old lady was 12 weeks pregnant with her second
pregnancy. She had a history of depressive illness. There
was also a history of marital discord. While abroad with her
family, she attempted suicide and overdosed on a cocktail of
antihypertensives (which belonged to her father-in-law),
anxiolytics and antidepressants. The patient was found
unconscious, and after admission to an intensive care unit
(ICU) a computer tomographic (CT) scan of her brain
showed gross ischaemic changes in the cerebral cortex
consistent with severe hypoxic brain damage.
After air transfer to our ICU, 4 weeks after the suicide
attempt, she had a Glasgow Coma Score (GCS) of 7
(spontaneous eye opening, no verbal response and an
extensor response to pain). She responded to distressing
situations by becoming increasingly agitated, sweaty and
tachypnoeic. A tracheostomy had been performed at the

referring hospital. The tracheostomy tube was changed to a
fenestrated type to facilitate spontaneous ventilation, an
enterostomy feeding tube was sited and a central venous
catheter (CVP) was inserted to aid ¯uid management. She
was treated for multiple respiratory, urinary and gastrointestinal infections with appropriate antibiotics, but continued to demonstrate intermittent pyrexia and persistent
tachycardia.
A fetal ultrasound scan showed that the fetus was
intact and small for date at 16 weeks, but did not
appear to have any anatomical abnormality. A plan to
manage the pregnancy expectantly was made. This
involved weekly clinical assessment supplemented by
regular ultrasound assessment of fetal growth, liquor
volume and umbilical artery waveforms. In the absence
of fetal or maternal indications, a requirement for early
delivery or Caesarean section was not anticipated. These
®ndings were discussed extensively with the family, who
were particularly concerned about fetal survival.
Arrangements were made initially to transfer the patient
to a rehabilitation unit once her condition had stabilized.
However, this decision was reversed after careful
consideration of the risks to the patient and the fetus.
She therefore remained in the ICU until delivery. Whilst
in the ICU she required substantial nursing care,
although she was not arti®cially ventilated.
The ICU nursing staff were instructed in the palpation of
uterine contractions and encouraged to request midwifery
support if in doubt. At 35 weeks gestation, the patient
developed regular uterine contractions, which seemed to be

Ó The Board of Management and Trustees of the British Journal of Anaesthesia 2000

Ayorinde et al.

causing considerable distress. A lumbar epidural catheter
was sited in the lateral position and an infusion of 0.25%
bupivacaine commenced at a rate of 6±10 ml h±1, which was
adjusted according to the patient's agitation and the
cardiovascular parameters. She was transferred to the
delivery suite after spontaneous rupture of the membranes
occurred. Continuous fetal heart rate monitoring identi®ed a
baseline rate of 140 beats min±1 with normal variability.
Taking into account the patient's previous normal
delivery and the small size of the fetus, it was decided to
allow a trial of labour and to anticipate vaginal delivery.
However, it was felt that there should be a low threshold for
operative delivery. As the patient was not in established
labour, the fetal heart rate decelerations, the need for
maternal analgesia and the suspected rupture of membranes
prompted a decision to induce labour by arti®cial membrane
rupture with oxytocin augmentation. The patient appeared
comfortable with the epidural solution being infused at a
rate of 8 ml h±1. Four hours after admission to the delivery
suite, she delivered a healthy 2.2 kg boy by spontaneous
vaginal delivery, who had an Apgar score of 10 at 5 min.
She was transferred to a rehabilitation unit 2 days later.

Discussion
This is the ®rst report, to our knowledge, of a pregnancy in
which the fetus survived maternal PVS during the entire
second and third trimesters, thus presenting a therapeutic
dilemma during the pregnancy and at delivery of the fetus.
PVS is a state of eyes-open unconsciousness with sleep±
wake cycles in which patients are incapable of awareness of
themselves or their environment.2 3 The use of the term PVS
usually implies that the vegetative state has continued for
more than 1 month.4 A patient in this state is not expected to
recover, although this is the subject of controversy. There
have been rare documented cases of late recovery of patients
who had been diagnosed as being in a PVS.5 Hence the
diagnosis should be considered carefully before it is made
because of the enormous ®nancial, emotional and social
implications of the management of these patients. Apart
from establishing lack of awareness, the cause and
irreversibility of the syndrome must be obvious.6 The
differential diagnosis includes prolonged coma (a state of
unconsciousness in which the eyes are closed and sleep±
wake cycles are absent) and the locked-in syndrome
(resulting from brainstem lesions which disrupt the voluntary control of movement without abolishing either arousal
or the content of awareness), both of which can also result
from severe brain injury.7
Patients with PVS usually breathe spontaneously,
although a tracheostomy may be necessary to ensure airway
protection. Patients may exhibit a bizarre variety of
neurological signs, including arousal to painful stimuli,
grimacing, limb movement, spontaneous head movements,
chewing, and grinding the teeth.4 Most of these features
were present in our patient, but with the additional feature

that she was pregnant, exposing her to the potential risk of
obstetric complications such as aortocaval compression,
pregnancy-induced hypertension and premature labour.
Termination of the pregnancy was not considered appropriate as the fetus had survived the hypoxic insult of a
suicide attempt and had a good chance of survival as long as
the mother was alive.
The choice of an appropriate place to manage the patient
was a contentious issue. Managing her in a side room in the
delivery suite would have been considered as a cheaper
alternative to high-dependency or intensive care. However,
this would not have provided optimal patient care in view of
the presence of a tracheostomy tube and the lack of
appropriate expertise in airway management among the
labour ward staff. The potential requirement for arti®cial
ventilation, the degree of nursing and physician input
required and the obstetric and fetal risks indicated that the
intensive care unit, rather than the high-dependency unit,
was the most appropriate place to optimize her management. Once the fetus had been delivered without any
maternal complications, the patient was deemed to present a
risk pro®le similar to a non-pregnant patient in a PVS, who
would usually be managed safely in a rehabilitation unit.
Assessment of pain and the provision of safe and effective
analgesia during labour and delivery in a patient in a PVS is
challenging because the patient cannot communicate.
Physiological parameters such as tachycardia, hypertension
and tachypnoea are commonly used as indicators of
inadequate analgesia, but all lack speci®city. The safety
and ef®cacy of epidural analgesia for the management of
labour pain is well documented.8 Despite the absence of
reports describing the use of epidural analgesia for a similar
case, we could not ®nd any justi®able reason for withholding epidural analgesia in our patient. We decided to use a
plain bupivacaine infusion to avoid the further neurological
depression that may have been induced by opioids. A
continuous infusion, rather than bolus doses of epidural
local anaesthetic, was chosen to minimize cardiovascular
compromise, especially since dermatomal sensory levels of
the block could not be assessed. The medical condition of
the patient, the fetal heart rate decelerations in the
intrapartum period and the anaesthetic considerations
indicated a low threshold for emergency Caesarean section.
Although regional anaesthesia is now established as the
preferred technique for Caesarean section,9 this requires a
dense sensory block extending from the fourth thoracic
dermatome to the ®fth sacral dermatome.10 However, we
could not establish the extent of the block in this patient.
Therefore, our anaesthetic plan in the event of emergency
Caesarean section was to administer a general anaesthetic.
Provision was made for a cuffed tracheostomy tube and
antacid prophylaxis, as for any obstetric patient undergoing
general anaesthesia. The epidural could also have been used
for postoperative analgesia.
In conclusion, this unusual case illustrates that pregnant
patients in a PVS may receive low-dose epidural analgesia

480

Pregnancy in persistent vegetative state

in labour, but general anaesthesia would be indicated for
Caesarean section. Nursing support, medical management,
fetal assessment and deciding the optimum time for delivery
are challenging decisions requiring a multidisciplinary
approach to this clinical situation.

References
1 Jennett B, Plum F. Persistent vegetative state after brain damage.
Lancet 1972; i: 734±7
2 Multi-Society Task Force on PVS. Medical aspects of the
persistent vegetative state (Part 1). N Engl J Med 1994; 330:
1499±508
3 Multi-Society Task Force on PVS. Medical aspects of the
persistent vegetative state (Part 2). N Engl J Med 1994; 330:
1572±9

481

4 Zeman A. Persistent vegetative state. Lancet 1997; 350: 795±9
5 Childs NL, Mercer WN. Late improvement in consciousness
after post-traumatic vegetative state. N Engl J Med 1996; 334:
24±5
6 Wade TW, Johnston C. The permanent vegetative state:
practical guidance on diagnosis and management. Br Med J
1999; 319: 841±4
7 Plum F, Posner JB. The diagnosis of stupor and coma.
Philadelphia: FADC, 1980
8 Yancey MK, Pierce B, Schweitzer D, Daniels D. Observations on
labor epidural analgesia and operative delivery rates. Am J Obstet
Gynecol 1999; 18: 353±9
9 May AE. Epidurals for Childbirth. Oxford: Oxford University Press,
1994
10 Morgan BM. `Walking' epidurals in labour. Anaesthesia 1995; 50:
839±40