Post-dural puncture headache: pathogenesis, prevention and treatment
British Journal of Anaesthesia, 2003, Vol.
91, No. 5 718-729
D. K. Turnbull*,1 and
D. B. Shepherd1,2
1 Academic Anaesthetic Unit, University of Sheffield, K Floor, Royal Hallamshire Hospital, and 2 Jessop Hospital for Women, Sheffield S10 2JF, UK
Corresponding author. E-mail: totleytiger@ yahoo.co. uk
Abstract
Spinal anaesthesia developed in the late 1800s with the work of Wynter, Quincke and Corning. However, it was the German surgeon, Karl August Bier in 1898, who probably gave the first spinal anaesthetic. Bier also gained first-hand experience of the disabling headache related to dural puncture. He correctly surmised that the headache was related to excessive loss of cerebrospinal fluid (CSF). In the last 50 yr, the development of fine-gauge spinal needles and needle tip modification, has enabled a significant reduction in the incidence of post-dural puncture headache. Though it is clear that reducing the size of the dural perforation reduces the loss of CSF, there are many areas regarding the pathogenesis, treatment and prevention of post-dural puncture headache that remain contentious. How does the microscopic pattern of collagen alignment in the spinal dura affect the dimensions of the dural perforation? How do needle design, size and orientation influence leakage of CSF through the dural perforation? Can pharmacological methods reduce the symptoms of post-dural puncture headache? By which mechanism does the epidural blood patch cure headache? Is there a role for the prophylactic epidural blood patch? Do epidural saline, dextran, opioids and tissue glues reduce the rate of CSF loss? This review considers these contentious aspects of post-dural puncture headache.
Br J Anaesth 2003; 91: 718–29
Keywords: anaesthetic techniques, subarachnoid; analeptics, caffeine; complications, dural puncture; complications, headache
History
Spinal anaesthesia developed in the late 1800s. In 1891, Wynter and Quincke95 aspirated cerebrospinal fluid (CSF) from the subarachnoid space for the treatment of raised intracranial hypertension associated with tuberculous meningitis. The catheters and trochars used were probably about 1 mm in diameter and would certainly have led to a post-dural puncture headache. However, all Quincke and Wynters’ subjects died soon after.
In 1895, John Corning, a New York physician specializing in diseases of the mind and nervous system, proposed that local anaesthesia of the spinal cord with cocaine may have therapeutic properties.50 Corning injected cocaine 110 mg at the level of the T11/12 interspace in a man to treat habitual masturbation. Despite being accredited with the first spinal anaesthetic, it is unlikely from his description and the dose of cocaine that his needle entered the subarachnoid space.82 In August 1898, Karl August Bier,137 a German surgeon, injected cocaine 10–15 mg into the subarachnoid space of seven patients, himself and his assistant, Hildebrandt. Bier, Hildebrandt and four of the subjects all described the symptoms associated with post-dural puncture headache. Bier surmised that the headache was attributable to loss of CSF. By the early 1900s, there were numerous reports in the medical literature of the application of spinal anaesthesia using large spinal needles.75 Headache was reported to be a complication in 50% of subjects. At that time, the headache was said to resolve within 24 h.
Ether anaesthesia was introduced into obstetric practice in 1847, shortly after Morton’s public demonstration. Despite the obvious advantages of regional anaesthesia for the relief of labour pain, it was not until a Swiss obstetrician in 1901 used intrathecal cocaine for the relief of pain in the second stage of labour that regional anaesthesia for obstetrics was popularized.49 Though both vomiting and a high incidence of post-dural puncture headache were noted, it was the high mortality rate in Caesarean deliveries performed under spinal anaesthesia (1 in 139) that led to the abandonment of this technique in the 1930s. The period from 1930 to 1950 has often been referred to as the ‘dark ages of obstetric anaesthesia’, when natural childbirth and psychoprophylaxis were encouraged.
In 1951, Whitacre and Hart59 developed the pencil-point needle, based on the observations of Greene53 in 1926. Developments in needle design since that time have led to a significant reduction in the incidence of post-dural puncture headache. However, dural puncture headache remains a disabling complication of needle insertion into the subarachnoid space.
Pathophysiology of dural puncture
Anatomy of the spinal dura mater
The spinal dura mater is a tube
extending from the foramen magnum to the second segment of the
sacrum. It contains the spinal cord and nerve roots that pierce it.
The dura mater is a dense, connective tissue layer made up of
collagen and elastic fibres. The classical description of the spinal
dura mater is of collagen fibres running in a longitudinal
direction.53
This had been supported by histological studies of the dura
mater.93
Clinical teaching based upon this view of the dura recommends that
a cutting spinal needle be orientated parallel rather than at
right angles to these longitudinal dural fibres. Orientating
the needle at right angles to the parallel fibres, it was said
would cut more fibres. The cut dural fibres, previously under
tension, would then tend to retract and increase the longitudinal
dimensions of the dural perforation, increasing the likelihood
of a post-spinal headache. Clinical studies had confirmed that
post-dural puncture headache was more likely when the cutting
spinal needle was orientated perpendicular to the direction of
the dural fibres. However, recent light and electron microscopic
studies of human dura mater have contested this classical description
of the anatomy of the dura mater.102
These studies describe the dura mater as consisting of collagen
fibres arranged in several layers parallel to the surface. Each layer
or lamellae consists of both collagen and elastic fibres that do not
demonstrate specific orientation.43
The outer or epidural surface may indeed have dural fibres arranged
in a longitudinal direction, but this pattern is not repeated through
successive dural layers. Recent measurements of dural thickness have
also demonstrated that the posterior dura varies in thickness, and
that the thickness of the dura at a particular spinal level is not
predictable within an individual or between individuals.102
Dural perforation in a thick area of dura may be less likely to lead
to a CSF leak than a perforation in a thin area, and may explain
the unpredictable consequences of a dural perforation.
Cerebrospinal fluid
CSF production occurs mainly in the choroid
plexus, but there is some evidence of extrachoroidal production.
About 500 ml of CSF is produced daily
(0.35 ml min–1). The CSF volume in the adult is
approximately 150 ml, of which half is within the cranial
cavity. The CSF pressure in the lumbar region in the horizontal
position is between 5 and 15 cm H2O. On assuming
the erect posture, this increases to over 40 cm H2O.
The pressure of the CSF in children rises with age, and may be
little more than a few cm H2O in early life.
Dura mater and response to trauma
The consequences of perforation
of the spinal or cranial dura are that there will be leakage of CSF.
Neurosurgical experience of dural perforation is that even minor
perforations need to be closed, either directly or through the
application of synthetic or biological dural graft material. Failure
to close the dural perforation may lead to adhesions, continuing CSF
leak, and the risk of infection. There are few experimental studies
of the response of the dura to perforation.70
In 1923, it was noted that deliberate dural defects in the cranial
dura of dogs took approximately one week to close. The closure was
facilitated through fibroblastic proliferation from the cut edge of
the dura. Work published in 195970
dismissed the notion that the fibroblastic proliferation arose from
the cut edge of the dura. This study maintained that the dural repair
was facilitated by fibroblastic proliferation from surrounding tissue
and blood clot. The study also noted that dural repair was promoted
by damage to the pia arachnoid, the underlying brain and the
presence of blood clot. It is therefore possible that a spinal
needle carefully placed in the subarachnoid space does not
promote dural healing, as trauma to adjacent tissue is minimal.
Indeed, the observation that blood promotes dural healing agrees
with Gormley’s original observation that bloody taps were less
likely to lead to a post-dural puncture headache as a consequence
of a persistence CSF leak.51
Needle tip deformation and dural perforation
It has been proposed
that contact with bone during insertion may lead to spinal needle tip
deformation.67
90
Damaged needle tips could lead to an increase in the size of the
subsequent dural perforation. Recent in vivo studies have
demonstrated that the cutting type spinal needle is more likely to be
deformed after bony contact than comparable sized pencil-point
needles.90
However, no in vivo67
or in vitro work has yet demonstrated an increase in the size
of dural perforation where damaged needles are used.
Consequences of dural puncture
Puncture of the dura has the
potential to allow the development of excessive leakage of CSF.
Excess loss of CSF leads to intracranial hypotension and a
demonstrable reduction in CSF volume.52
After the development of post-dural puncture headache, the
presence of a CSF leak has been confirmed with radionuclide
cisternography,100
radionuclide myelography, manometric studies, epiduroscopy and
direct visualization at laminectomy. The adult subarachnoid
pressure of 5–15 cm H20 is reduced to
4.0 cm H20 or less.100
The rate of CSF loss through the dural perforation29
(0.084–4.5 ml s–1) is generally greater
than the rate of CSF production (0.35 ml min–1),
particularly with needle sizes larger than 25G.29
101
Gadolinium-enhanced MRI, in the presence of a post-dural puncture headache, frequently demonstrates ‘sagging’ of the intracranial structures. The MRI may or may not demonstrate meningeal enhancement.56 The meningeal enhancement is attributable to vasodilatation of thin-walled vessels in response to the intracranial hypotension. Histological studies have confirmed that the vasodilation of meningeal vessels is unrelated to an inflammatory response.56
Although the loss of CSF and lowering of CSF pressure is not disputed, the actual mechanism producing the headache is unclear. There are two possible explanations. First, the lowering of CSF pressure causes traction on the intracranial structures in the upright position. These structures are pain sensitive, leading to the characteristic headache. Secondly, the loss of CSF produces a compensatory venodilatation vis-à-vis the Monro–Kellie doctrine.52 The Monro–Kellie doctrine, or hypothesis, states that the sum of volumes of the brain, CSF, and intracranial blood is constant. The consequence of a decrease in CSF volume is a compensatory increase in blood volume. The venodilatation is then responsible for the headache.
Incidence
The incidence of post-dural puncture headache was 66% in 1898.137 This alarmingly high incidence of post-spinal headache was likely attributable to the use of large gauge, medium bevel, cutting spinal needles (needles 5, 6 and 7, Fig. 1). In 1956, with the introduction of 22G and 24G needles, the incidence was estimated to be 11%.132
|
Today the use of fine gauge pencil-point needles, such as the Whitacre and Sprotte® has produced a greater reduction in the incidence of post-dural puncture headache, which varies with the type of procedure and patients involved. It is related to the size and design of the spinal needle used (Fig. 1; Table 1),36 the experience of the personnel performing the dural puncture,35 and the age and sex of the patient.
|
Spinal anaesthesia
Anaesthetists have been active in attempting to reduce the incidence of post-spinal headache. Reducing the size of the spinal needle has made a significant impact on the incidence of post-spinal headache. The incidence is
40% with a 22G needle;
25% with a 25G needle;4
44
2%–12% with a 26G Quincke needle;4
45
and <2% with a 29G needle.47
However, technical difficulties leading to failure of the spinal
anaesthetic are common with needles of 29G or smaller.47
In 1951, Whitacre and Hart59
introduced the ‘atraumatic’ spinal needle (needle 3, Fig. 1).
This design offered the handling characteristics of larger needles
with a low incidence of post-spinal headache (Table 1).
Needle modifications since that time, such as the
Sprotte®119
and Atraucan®63
needles, promise further reductions in post-spinal headache.
Diagnostic lumbar puncture
The acceptance of small gauge needles
for diagnostic lumbar puncture has been slow to develop. Until
recently, diagnostic lumbar puncture was commonly performed with a
20G or even 18G medium bevel cutting needle with a high incidence of
post-spinal headache. A recent publication promoted the virtues of a
20G needle for reducing the incidence of dural puncture
headache!125
Though anaesthetists are in general critical of the use of large
gauge needles for lumbar puncture,105
neurologists maintain that adequate flow of CSF can only be achieved
with spinal needles of 22G or greater.18
Obstetrics
The parturient is at particular risk of dural puncture
and the subsequent headache because of their sex, young age, and
the widespread application of epidural anaesthesia.44
In parturients receiving epidural anaesthesia, the incidence of dural
puncture is between 0 and 2.6%.104
The incidence is inversely related to the experience of the
anaesthetist,80
and is said to be reduced by orientation of the needle bevel parallel
to the dural fibres.87
Loss of resistance to air confers a higher risk of dural puncture
than loss of resistance to fluid.105
After a dural puncture with a 16G Tuohy needle, up to 70% of subjects
will report symptoms related to low CSF pressure.26
Despite the high incidence of headache consequent upon dural puncture
with a Tuohy needle, the anaesthetist needs to consider a
differential diagnosis, as intracranial haematoma,65
or tumour38
presenting with similar symptoms to, or in association with, a
post-dural puncture headache have been described.
In the presence of a known dural puncture, it is often recommended that pushing in the second stage should be avoided.88 The evidence to support this assertion is far from conclusive, and anger from the parturient about the medicalization of her labour is best avoided.26 133
Children
Post-dural puncture headache is reported as uncommon in
children.14
Although low CSF pressure or other physiological differences
have been proffered as reasons to explain the low incidence in
children, it is likely that a low reporting rate is the explanation.
Groups that have explored the incidence of post-spinal headaches
in children have found rates comparable to young adults.73
Prevention
Spinal needles have undergone numerous modifications in recent years, the aim being to reduce the incidence of dural puncture headache. The principal factor responsible for the development of a dural puncture headache is the size of the dural perforation. Other factors such as the shape of the dural perforation and the orientation of the spinal needle have a less significant role.
Needle size
Large spinal needles will clearly produce large dural
perforations where the likelihood of a dural puncture headache is
high. Conversely, the smaller needles produce small dural
perforations with a lower incidence of headache. Fine gauge spinal
needles, 29G or smaller, are technically more difficult to
use,64
and for spinal anaesthesia at least, are associated with a high
failure rate.45
A balance has to be struck between the risks of dural puncture
headache and technical failure. 25G, 26G and 27G69
needles probably represent the optimum needle size for spinal
anaesthesia. Neurologists argue that for the purposes of aspiration
of CSF and measurement of CSF pressure, 22G needles are the
smallest practical needles.
Needle orientation
There are many clinical,79
87
and laboratory,36
101
studies that lend credence to the hypothesis that perpendicular
orientation of the bevel of a spinal or epidural needle leads to a
reduction in the incidence of post-dural puncture headache.
Needle design
Over the years since Quincke and Bier, a large
number of needle designs have been introduced. The Quincke type is
the standard needle with a medium cutting bevel and the orifice at
the needle tip (needle 7, Fig. 1).
In 1926, Greene53
proposed a needle tip design with a non-cutting edge that would
separate the dural fibres to avoid post-dural puncture headache. In
1951, the Whitacre needle was introduced and, in 1987, the Sprotte
needle. The generic term for these needles is pencil-point or
atraumatic, though in truth they are neither. The Whitacre needle
(needle 3, Fig. 1)
has a diamond shaped tip, and the Sprotte needle (needle 2, Fig. 1)
tip is conical. The orifice is up to 0.5 mm from the needle tip.
Clinical and laboratory29
studies have confirmed that pencil-point needles produce fewer
post-dural puncture headaches than medium bevel cutting needles.
However, there are disadvantages. Paraesthesia has been observed
with the pencil-point needles.115
The reason may lie in the distance from the tip of the needle to the
orifice. The tip has to be passed at least 0.5 mm into the
subarachnoid space before the orifice enters the subarachnoid space.
The tip then has the opportunity to impinge upon the stretched cauda
equina. Giving credence to this hypothesis, paraesthesia is
uncommon with the short bevel needles or the Atraucan®
needle.115
The problem of low CSF flow and paraesthesia seen with the pencil-point needles has promoted the search for novel needle designs. The Atraucan® (needle 1, Fig. 1) has recently been marketed. It has an orifice at the tip of the needle. The Atraucan® has a narrow cutting tip and an atraumatic bevel. Initial reports of these needles are promising as regards ease of use and low dural puncture headache rate.115
Operator skill level and fatigue
It has been suggested that the
incidence of inadvertent dural puncture during epidural anaesthesia
is inversely related to operator experience.104
However, sleep deprivation, operator fatigue and the effect of night
work may be a confounding variable producing the higher incidence of
inadvertent dural puncture in junior personnel performing epidural
analgesia.
Presentation of dural puncture headache
Onset
Headache and backache are the dominant symptoms that
develop after accidental dural puncture. Ninety per cent of
headaches will occur within 3 days of the procedure,104
and 66% start within the first 48 h.76
Rarely, the headache develops between 5 and 14 days after the
procedure. Headache may present immediately after dural
puncture.133
However, this is rare, and its occurrence should alert the physician
to alternative causes.
Symptoms
Headache is the predominant, but not ubiquitous
presenting complaint.83
The headache is described as severe, ‘searing and spreading
like hot metal’.133
The common distribution is over the frontal and occipital areas
radiating to the neck and shoulders. The temporal, vertex and nuchal
areas are reported less commonly as the site of discomfort, although
neck stiffness may be present. The pain is exacerbated by head
movement, and adoption of the upright posture, and relieved by lying
down. An increase in severity of the headache on standing is the
sine qua non of post-dural puncture headache.
Other symptoms associated with dural puncture headache include nausea, vomiting, hearing loss,78 tinnitus, vertigo, dizziness and paraesthesia of the scalp, and upper108 and lower limb pain. Visual disturbances such as diplopia or cortical blindness have been reported.132 Cranial nerve palsies are not uncommon.16 Two cases of thoracic back pain without headache have been described.37 Neurological symptoms may precede the onset of grand mal seizures. Intracranial subdural haematomas, cerebral herniation and death,39 have been described as a consequence of dural puncture. Unless a headache with postural features is present, the diagnosis of post-dural puncture headache should be questioned, as other serious intracranial causes for headache must be excluded.3
Diagnosis
The history of accidental or deliberate dural puncture
and symptoms of a postural headache, neck ache and the presence of
neurological signs, usually guide the diagnosis. Where there is doubt
regarding the diagnosis of post-dural puncture headache, additional
tests may confirm the clinical findings. A diagnostic lumbar
puncture may demonstrate a low CSF opening pressure or a ‘dry
tap’, a slightly raised CSF protein, and a rise in CSF
lymphocyte count. An MRI may demonstrate: diffuse dural enhancement,
with evidence of a sagging brain; descent of the brain, optic
chiasm, and brain stem; obliteration of the basilar cisterns; and
enlargement of the pituitary gland.85
CT myelography, retrograde radionuclide myelography, cisternography,
or thin section MRI130
can be used to locate the spinal source of the CSF leak.
Differential diagnosis
The diagnosis of post-dural puncture
headache is frequently clear from the history of dural puncture and
the presence of a severe postural headache. However, it is important
to consider alternative diagnoses (Table 2)
as serious intracranial pathology may masquerade as a post-dural
puncture headache. Clinicians should remember that intracranial
hypotension can lead to intracranial haemorrhage through tearing of
bridging dural veins,65
94
and a delay in diagnosis and treatment can be dangerous.
Diagnoses that may masquerade as post-dural puncture headache
include intracranial tumours,3
38
intracranial haematoma,32
40
pituitary apoplexy,77
cerebral venous thrombosis,122
134
migraine, chemical or infective meningitis,106
and non-specific headache. It has been estimated that 39% of
parturients report symptoms of a headache unrelated to dural puncture
following delivery.120
|
Duration
The largest follow-up of post-dural puncture headache is still that of Vandam and Dripps in 1956.132 They reported that 72% of headaches resolved within 7 days, and 87% had resolved in 6 months (Table 3). The duration of the headache has remained unchanged since that reported in 1956.26 In a minority of patients the headache can persist.133 Indeed, case reports have described the persistence of headache for as long as 1–8 yr after dural puncture.80 It is interesting to note that even post-dural puncture headaches of this duration have been successfully treated with an epidural blood patch.72
|
Treatment
Overview
The literature regarding the treatment of post-dural
puncture headache often involves small numbers of patients, or uses
inappropriate statistical analysis. Studies observing the effects of
treatments in post-dural puncture headache often fail to recognize
that, with no treatment, over 85% of post-dural puncture
headaches will resolve within 6 weeks (Table 3).
Psychological
Patients who develop post-dural puncture headache
may reveal a wide range of emotional responses from misery and tears
to anger and panic. It is important both from a clinical and
medico-legal point of view, to discuss the possibility of headache
before a procedure is undertaken that has a risk of this
complication. Even so, this discussion will not prepare the patient
for the sensations he or she feels should the headache
develop.133
Obstetric patients are particularly unfortunate should they develop
this complication, as they expect to feel well and happy and to
be able to look after their new baby. It is important to give
the mother a thorough explanation of the reason for the
headache, the expected time course, and the therapeutic options
available. Regular review is essential to monitor the course and
therapeutic manoeuvres undertaken.
Simple
Bed rest has been shown to be of no benefit.118
Supportive therapy such as rehydration, acetaminophen, non-steroidal
anti-inflammatory drugs, opioids, and antiemetics may control the
symptoms and so reduce the need for more aggressive therapy,89
but do not provide complete relief.44
Posture
If a patient develops a headache, they should be
encouraged to lie in a comfortable position. The patient will often
have identified this, without the intervention of an
anaesthetist. There is no clinical evidence to support the
maintenance of the supine position before or after the onset of the
headache as a means of treatment.68
The prone position has been advocated, but it is not a comfortable
position for the post-partum patient. The prone position raises the
intra-abdominal pressure, which is transmitted to the epidural space
and may alleviate the headache. A clinical trial of the prone
position following dural puncture failed to demonstrate a reduction
in post-dural puncture headache.55
Abdominal binder
A tight abdominal binder raises the
intra-abdominal pressure. The elevated intra-abdominal pressure is
transmitted to the epidural space and may relieve the headache.
Unfortunately, tight binders are uncomfortable and are seldom used in
current practice. There are few units that would recommend this
approach.86
Pharmacological treatment
The aim of management of post-dural
puncture headache is to: (i) replace the lost CSF; (ii) seal the
puncture site; and (iii) control the cerebral vasodilatation.
A number of therapeutic agents have been suggested for the management of post-dural puncture headache. The main problem in choosing the most appropriate one is the lack of large, randomized, controlled clinical trials.
DDAVP, ACTH
A report in 1964 identified 49 methods for treating
post-spinal headache.127
There appears to be no limit to the imagination of physicians in
treatments offered for post-spinal headache. However, there is a lack
of statistical data to support their ideas. Regarding DDAVP
(desmopressin acetate), intramuscular administration before lumbar
puncture was not shown to reduce the incidence of post-dural puncture
headache.57
ACTH (adrenocorticotroph ic hormone)21
has been administered as an infusion (1.5 µg kg–1),
but inadequate statistical analysis prevents assessment of the
value of ACTH.
Caffeine
Caffeine is a central nervous system stimulant that
amongst other properties produces cerebral vasoconstriction. I.V.
caffeine 0.5 g was recommended as a treatment of post-dural
puncture headache in 1944.62
It is available in an oral and i.v. form. The oral form is well
absorbed with peak levels reached in 30 min. Caffeine crosses
the blood–brain barrier and the long half-life of 3–7.5 h allows
for infrequent dosing schedules.
The most frequently quoted work on the treatment of post-dural puncture headache with caffeine is that of Sechzer.113 114 He evaluated the effects of one or two 0.5 g doses of i.v. caffeine on subjects with established post-dural puncture headache. There are some statistical and methodological flaws in this study, but it was concluded that i.v. caffeine is an effective therapy for post-dural puncture headache.
Dose
The dose now recommended for
the treatment of post-dural puncture headache is 300–500 mg of
oral or i.v. caffeine once or twice daily.12
66
One cup of coffee contains about 50–100 mg of caffeine and soft
drinks contain 35–50 mg. The LD50 for caffeine is of
the order of 150 mg kg–1. However, therapeutic
doses have been associated with central nervous system
toxicity,9
and atrial fibrillation.
Mode of action
It is assumed that
caffeine acts through vasoconstriction of dilated cerebral
vessels.12
If cerebral vasodilatation were the source of the pain, cerebral
vasoconstriction might limit the pain experienced. Indeed, it has
been demonstrated that caffeine causes a reduction in cerebral blood
flow,116
but this effect is not sustained. Caffeine therapy is simple to
administer compared with the technical skills required to perform an
epidural blood patch. Were caffeine as successful as suggested by
previous reports, it would no doubt be widely advocated. However, a
North American hospital survey of the treatment of post-dural
puncture headache identified that most hospital practitioners had
abandoned the use of caffeine as they had found it
ineffective.8
The effects of caffeine on post-dural puncture headache seem, at
best, temporary.12
In addition, caffeine is not a therapy without complications,9
and does not restore normal CSF dynamics, thus leaving the patient
at risk from the serious complications associated with low CSF
pressure.
Sumatriptan
The treatment for migranous headaches has focused on
modification of cerebral vascular tone. Sumatriptan is a
5-HT1D receptor agonist that promotes cerebral
vasoconstriction, in a similar way to caffeine.123
Sumatriptan is advocated for the management of migraine and recently,
for post-dural puncture headache. There have been only a few case
reports where sumatriptan was used successfully to manage post-dural
puncture headache.61
However, a recent controlled trial found no evidence of benefit
from Sumatriptan for the conservative management of post-dural
puncture headache.23
Epidural blood patch
History
After the observation that ‘bloody
taps’ were associated with a reduced headache rate,51
the concept of the epidural blood patch has developed. The theory is
that the blood, once introduced into the epidural space, will clot
and occlude the perforation, preventing further CSF leak. The high
success rate and the low incidence of complications have established
the epidural blood patch as the standard against which to
evaluate alternative methods to treat post-dural puncture
headache.
Technique
The presence of fever,
infection on the back, coagulopathy, or patient refusal are
contraindications to the performance of an epidural blood
patch.1
As a precautionary measure, a sample of the subject’s blood should be
sent to microbiology for culture.27
With the patient in the lateral position, the epidural space is
located with a Tuohy needle at the level of the supposed dural
puncture or an intervertertebral space lower. The operator should be
prepared for the presence of CSF within the epidural space. Up to
30 ml of blood is then taken from the patient’s arm and
injecting slowly through the Tuohy needle. Should the patient
describe lancinating pain of dermatomal origin the procedure must be
stopped.27
There is no consensus as to the precise volume of blood required.
Most practitioners now recognise that the 2–3 ml of blood
originally described by Gormley is inadequate, and that 20–30 ml
of blood is more likely to guarantee success.27
Larger volumes, up to 60 ml,97
have been used successfully in cases of spontaneous intracranial
hypotension. At the conclusion of the procedure, the patient is asked
to lie still for one1
33
or, preferably, 2 h,81
and is then allowed to walk.
Contraindications
Contraindications
include those that normally apply to epidurals, but include a raised
white cell count, pyrexia and technical difficulties. Limited
experience with HIV-positive patients suggest that it is acceptable
providing no other bacterial or viral illnesses are active.126
Epidural blood patch following diagnostic lumbar puncture in the
oncology patient raises the potential for seeding the neuroaxis with
neoplastic cells. One case has been reported of a successful patch
without complications,109
and one case11
where the risks of central nervous system (CNS) seeding of leukaemia
were considered to outweigh the benefits of an epidural blood
patch.
The blood patch
Using either
radiolabelled red cells124
or an MRI scan,7
several studies have reported the degree of spread of the epidural
blood patch. After injection, blood is distributed caudally and
cephalad regardless of the direction of the bevel of the Tuohy
needle. The blood also passes circumferentially around to the
anterior epidural space. The thecal space is compressed and
displaced by the blood. In addition, the blood passes out of the
intervertebral foramina and into the paravertebral space. The mean
spread of 14 ml of blood is six spinal segments cephalad and
three segments caudad. Compression of the thecal space for the
first 3 h, and a presumed elevation of subarachnoid pressure,
may explain the rapid resolution of the headache. Compression of
the thecal sac is not, however, sustained and maintenance of
the therapeutic effect is likely to be attributable to the presence
of the clot eliminating the CSF leak. It has been observed that
CSF acts as a procoagulant, accelerating the clotting process.24
At 7–13 h, there is clot resolution leaving a thick layer
of mature clot over the dorsal part of the thecal sac. Animal studies
have demonstrated that 7 days after the administration of an epidural
blood patch, there is widespread fibroblastic activity and collagen
formation.34
74
Fortunately, the presence of blood does not initiate an inflammatory
process and there is no evidence of axonal oedema, necrosis or
demyelination.
Outcome
The technique has a success
rate of 70–98% if carried out more than 24 h after the dural
puncture.1
If an epidural blood patch fails to resolve the headache, repeating
the blood patch has a similar success rate. Failure of the second
patch and repeating the patch for a third or fourth time has
been reported. However, in the presence of persistent severe
headache, an alternative cause should be considered.
Complications
Immediate exacerbation
of symptoms and radicular pain have been described.136
These symptoms do not persist and resolve with the administration of
simple painkillers. Long-term complications of epidural blood patch
are rare. A single case report of an inadvertent subdural epidural
blood patch described non-postural, persistent headache and lower
extremity discomfort.103
The issue of the effect of the blood patch on the success of subsequent epidurals has been addressed.2 60 Though case reports describe limited spread of epidural analgesia99 after previous epidural blood patch, a large retrospective study over a 12-yr period60 found that subsequent epidural analgesia was successful in >96% of patients.
Prophylactic epidural blood
patch
Where the known incidence of post-dural puncture
headache is high, such as in the parturient, the use of a
prophylactic epidural blood patch after accidental dural puncture,
that is blood patching before the onset of symptoms, is an attractive
option. Prophylactic patching has generally been dismissed as
ineffective, but the evidence is conflicting. A controlled trial in
post-myelogram headaches,54
and one after spinal anaesthesia and after unintentional dural
puncture with an epidural needle,22
have confirmed the benefit of prophylactic patching. Those studies
that have not supported the use of prophylactic patching may have
used insufficient blood for the patch.22
The pressure gradient between the thecal and epidural space may be
high immediately after dural puncture and lead to patch separation
from the site of the perforation. Blood patching at that time may
therefore need a greater volume of blood to produce a successful
patch compared with a late patch, where the CSF pressure may be
lower.
Chronic headache
Patients may
present with features of a post-spinal headache never having received
an epidural or spinal injection. A report of six such cases, with
headaches that had been present between 1 and 20 yr, showed complete
relief of headache following lumbar epidural blood patch.91
It is interesting to speculate that these headaches may have been
attributable to unidentified spontaneous intracranial
hypotension.
Epidural saline
Concerns have been expressed about the potential
danger of an autologous epidural blood patch for the treatment of
post-dural puncture headache. The immediate resolution of the
headache with a blood patch is attributable to thecal compression
raising the CSF pressure. An epidural injection of saline would,
in theory, produce the same mass effect, and restore normal CSF
dynamics. As saline is a relatively inert and sterile solution,
epidural saline bolus or infusion appears to be an attractive
alternative. Regimens that have been advocated include: (i)
1.0–1.5 litre of epidural Hartmanns solution over
24 h, starting on the first day after dural puncture;28
84
121
(ii) up to 35 ml h–1 of epidural saline or
Hartmanns solution for 24–48 h, or after development of the
headache; (iii) a single 30 ml bolus of epidural saline after
development of headache;5
84
and (iv) 10–120 ml of saline injected as a bolus via the caudal
epidural space.6
129
Advocates of an epidural saline bolus or infusion maintain that the lumbar injection of saline raises epidural and intrathecal pressure. Reduction in the leak would allow the dura to repair. However, observations of the pressures produced in the subarachnoid and epidural space show that, despite a large rise in epidural pressure, the consequent rise in subarachnoid pressure maintains the differential pressure across the dura. The pressure rise is also not sustained and is dissipated within 10 min.129 The saline may induce an inflammatory reaction within the epidural space, promoting closure of the dural perforation. Histological studies have not demonstrated an inflammatory response following epidural Dextran 40 administration, however, in contrast to an autologous blood patch.74 There is no reason to suppose that epidural saline is more likely to accelerate dural healing through a proinflammatory action than Dextran 40. Thus, there are no studies that are able to demonstrate either a sustained rise in CSF pressure or accelerated closure of the dural perforation after the administration of epidural saline. Whilst there are many case reports describing the success of epidural saline, comparative trials with epidural blood patches have not demonstrated the long-term efficacy of epidural saline placement.5 It is difficult to conclude from the evidence therefore, that epidural saline administration will restore normal CSF dynamics. The administration of large volumes of epidural saline may result in intraocular haemorrhages through a precipitous rise in intracranial pressure.19
Epidural dextran
Despite the paucity of evidence to support
epidural saline, some observers have considered the epidural
administration of Dextran 40.117
Those studies that recommend Dextran 40, either as an infusion or as
a bolus, conclude that the high molecular weight and viscosity of
Dextran 40 slows its removal from the epidural space. The sustained
tamponade around the dural perforation allows spontaneous closure.
However, it is unlikely that Dextran 40 will act any differently to
saline in the epidural space. Any pressure rise within the
subarachnoid space would, like saline, be only transient.
Histological inspection of the epidural space after administration of
Dextran 40,74
does not demonstrate any inflammatory response that would promote the
healing process. The evidence for the administration of epidural
Dextran to treat post-dural puncture headache is not proven and the
theoretical argument to justify its use is poor.
Epidural, intrathecal and parenteral opioids
A number of authors
have advocated the use of epidural,42
intrathecal20
or parenteral morphine;41
the majority of these reports are either case reports or inadequately
controlled trials. Some of the studies used epidural morphine after
the onset of headache, others used epidural or intrathecal morphine
as prophylaxis or in combination with an intrathecal catheter.20
A controlled trial of intrathecal fentanyl as prophylaxis found no
evidence of a reduction in the incidence of post-spinal headache
after dural puncture with a 25-gauge spinal needle.31
Fibrin glue
Alternative agents to blood, such as fibrinous glue,
have been proposed to repair spinal dural perforations.48
Cranial dural perforations are frequently repaired successfully with
it. In the case of lumbar dural perforation, the fibrin glue may
be placed blindly or using CT-guided percutaneous injection.92
There is, however, a risk of the development of aseptic meningitis
with this procedure.111
In addition, manufacturers have recently warned against the
application of some types of tissue glue where it may be exposed to
nervous tissue.110
Intrathecal catheters
After accidental dural perforation with a
Tuohy needle, it has been suggested that placement of a spinal
catheter through the perforation may provoke an inflammatory reaction
that will seal the hole. Evidence to support this claim is
conflicting.30
135
The mean age of the patients in some of the trials has been
>50 yr, where the rate of post-dural puncture headache
is low. Some trials have used spinal microcatheters, 26G–32G;
others have placed 20G epidural catheters through an 18G Tuohy
needle.
Histopathological studies in animals and humans with long-term intrathecal catheters confirm the presence of an inflammatory reaction at the site of the catheter. Comparison between the effects of a catheter left in situ for 24 h and for several days or weeks would seem inappropriate.96 If, after accidental dural puncture with a Tuohy needle, the insertion of an intrathecal catheter reduced the post-dural puncture headache rate, then it would be worth considering. However, neurological complications, such as cauda equina syndrome and infection, should preclude the use of intrathecal catheters.
Surgery
There are case reports of persistent CSF leaks, that are
unresponsive to other therapies, being treated successfully by
surgical closure of the dural perforation.58
This is clearly a last resort treatment.
Conclusions
Post-dural puncture headache is a complication that should not to be treated lightly. There is the potential for considerable morbidity,10 even death.39 104 In the majority of cases, the problem will resolve spontaneously. In some patients, the headache lasts for months or even years.
Therapies that have been offered have not always arisen through the application of logic or reasoning. Gormley’s observation that bloody taps were less likely to give rise to headaches, though probably incorrect,71 has led to the widespread application of blood patching for the treatment of post-dural puncture headache. The benefit of prophylactic blood patching is not so clear but deserves consideration in those most at risk from a headache, such as the parturient, and after accidental dural perforation with a Tuohy needle. There are occasions when blood patches appear to be ineffective in treating the headache. It is wise to consider other causes of the headache before applying alternative therapeutic options. Surgical closure of the dural tear is an option of last resort.
References
1 Abouleish E, Vega S, Blendinger I, Tio TO. Long-term follow-up of epidural blood patch. Anesth Analg 1975; 54: 459–63[Abstract]
2 Abouleish E, Wadhwa RK, de la Vega S, Tan RN Jr, Lim Uy N. Regional analgesia following epidural blood patch. Anesth Analg 1975; 54: 634–6[Abstract]
3 Alfery DD, Marsh ML, Shapiro HM. Post-spinal headache or intracranial tumor after obstetric anesthesia. Anesthesiology 1979; 51: 92–4[CrossRef][ISI][Medline]
4 Barker P. Headache after dural puncture. Anaesthesia 1989; 44: 696–7[ISI][Medline]
5 Bart AJ, Wheeler AS. Comparison of epidural saline placement and epidural blood placement in the treatment of post-lumbar- puncture headache. Anesthesiology 1978; 48: 221–3[CrossRef][ISI][Medline]
6 Baysinger CL, Menk EJ, Harte E, Middaugh R. The successful treatment of dural puncture headache after failed epidural blood patch. Anesth Analg 1986; 65: 1242–4[ISI][Medline]
7 Beards SC, Jackson A, Griffiths AG, Horsman EL. Magnetic resonance imaging of extradural blood patches: appearances from 30 min to 18 h. Br J Anaesth 1993; 71: 182–8 [Abstract/Free Full Text]
8 Berger CW, Crosby ET, Grodecki W. North American survey of the management of dural puncture occurring during labour epidural analgesia. Can J Anaesth 1998; 45: 110–14 [Abstract/Free Full Text]
9 Bolton VE, Leicht CH, Scanlon TS. Postpartum seizure after epidural blood patch and intravenous caffeine sodium benzoate. Anesthesiology 1989; 70: 146–9[CrossRef][ISI][Medline]
10 Borum SE, Naul LG, McLeskey CH. Postpartum dural venous sinus thrombosis after postdural puncture headache and epidural blood patch. Anesthesiology 1997; 86: 487–90[CrossRef][ISI][Medline]
11 Bucklin BA, Tinker JH, Smith CV. Clinical dilemma: a patient with postdural puncture headache and acute leukemia. Anesth Analg 1999; 88: 166–7 [Free Full Text]
12 Camann WR, Murray RS, Mushlin PS, Lambert DH. Effects of oral caffeine on postdural puncture headache. A double-blind, placebo-controlled trial. Anesth Analg 1990; 70: 181–4[Abstract]
13 Campbell DC, Douglas MJ, Pavy TJ, Merrick P, Flanagan ML, McMorland GH. Comparison of the 25-gauge Whitacre with the 24-gauge Sprotte spinal needle for elective Caesarean section: cost implications. Can J Anaesth 1993; 40: 1131–5 [Abstract/Free Full Text]
14 Carbajal R, Simon N, Olivier-Martin M. Post-lumbar puncture headache in children. Treatment with epidural autologous blood (blood patch). Arch Pediatr 1998; 5: 149–52[CrossRef][ISI][Medline]
15 Carp H, Singh PJ, Vadhera R, Jayaram A. Effects of the serotonin-receptor agonist sumatriptan on postdural puncture headache: report of six cases. Anesth Analg 1994; 79: 180–2[ISI][Medline]
16 Carrero EJ, Agusti M, Fabregas N, Valldeoriola F, Fernandez C. Unilateral trigeminal and facial nerve palsies associated with epidural analgesia in labour. Can J Anaesth 1998; 45: 893–7 [Abstract/Free Full Text]
17 Carrie LE. Whitacre and pencil-point needles: some points to consider. Anaesthesia 1990; 45: 1097–8[ISI][Medline]
18 Carson D, Serpell M. Choosing the best needle for diagnostic lumbar puncture. Neurology 1996; 47: 33–7[Abstract]
19 Clark CJ, Whitwell J. Intraocular haemorrahge after epidural injection. Br Med J 1961; I: 1612–13
20 Cohen S, Amar D, Pantuck EJ, Singer N, Divon M. Decreased incidence of headache after accidental dural puncture in caesarean delivery patients receiving continuous postoperative intrathecal analgesia. Acta Anaesthesiol Scand 1994; 38: 716–18[ISI][Medline]
21 Collier BB. Treatment for post-dural puncture headache. Br J Anaesth 1994; 72: 366–7 [Free Full Text]
22 Colonna-Romano P, Shapiro BE. Unintentional dural puncture and prophylactic epidural blood patch in obstetrics. Anesth Analg 1989; 69: 522–3[ISI][Medline]
23 Connelly NR, Parker RK, Rahimi A, Gibson CS. Sumatriptan in patients with postdural puncture headache. Headache 2000; 40: 316–19[CrossRef][ISI][Medline]
24 Cook MA, Watkins-Pitchford JM. Epidural blood patch: a rapid coagulation response. Anesth Analg 1990; 70: 567–8[ISI][Medline]
25 Corbey MP, Bach AB, Lech K, Frorup AM. Grading of severity of postdural puncture headache after 27-gauge Quincke and Whitacre needles. Acta Anaesthesiol Scand 1997; 41: 779–84[ISI][Medline]
26 Costigan SN, Sprigge JS. Dural puncture: the patients’ perspective. A patient survey of cases at a DGH maternity unit 1983–1993. Acta Anaesthesiol Scand 1996; 40: 710–14[ISI][Medline]
27 Crawford JS. Experiences with epidural blood patch. Anaesthesia 1980; 35: 513–15[ISI][Medline]
28 Crawford JS. The prevention of headache consequent upon dural puncture. Br J Anaesth 1972; 44: 598–600 [Abstract/Free Full Text]
29 Cruickshank RH, Hopkinson JM. Fluid flow through dural puncture sites. An in vitro comparison of needle point types. Anaesthesia 1989; 44: 415–18[ISI][Medline]
30 Denny N, Masters R, Pearson D, Read J, Sihota M, Selander D. Postdural puncture headache after continuous spinal anesthesia. Anesth Analg 1987; 66: 791–4[ISI][Medline]
31 Devcic A, Sprung J, Patel S, Kettler R, Maitra-D’Cruze A. PDPH in obstetric anesthesia: comparison of 24-gauge Sprotte and 25-gauge Quincke needles and effect of subarachnoid administration of fentanyl. Reg Anesth 1993; 18: 222–5[ISI][Medline]
32 Diemunsch P, Balabaud VP, Petiau C, et al. Bilateral subdural hematoma following epidural anesthesia. Can J Anaesth 1998; 45: 328–31 [Abstract/Free Full Text]
33 DiGiovanni AJ, Dunbar BS. Epidural injections of autologous blood for postlumbar-puncture headache. Anesth Analg 1970; 49: 268–71[Medline]
34 DiGiovanni AJ, Galbert MW, Wahle WM. Epidural injection of autologous blood for postlumbar-puncture headache. II. Additional clinical experiences and laboratory investigation. Anesth Analg 1972; 51: 226–32[Medline]
35 Dittmann M, Schaefer HG, Renkl F, Greve I. Spinal anaesthesia with 29 gauge Quincke point needles and post-dural puncture headache in 2378 patients. Acta Anaesthesiol Scand 1994; 38: 691–3[ISI][Medline]
36 Dittmann M, Schafer HG, Ulrich J, Bond-Taylor W. Anatomical re-evaluation of lumbar dura mater with regard to postspinal headache. Effect of dural puncture. Anaesthesia 1988; 43: 635–7[ISI][Medline]
37 Dunbar SA, Katz NP. Post-dural puncture thoracic pain without headache: relief with epidural blood patch. Can J Anaesth 1995; 42: 221–3 [Abstract/Free Full Text]
38 Dutton DA. A ‘postspinal headache’ associated with incidental intracranial pathology. Anaesthesia 1991; 46: 1044–6[ISI][Medline]
39 Eerola M, Kaukinen L, Kaukinen S. Fatal brain lesion following spinal anaesthesia. Report of a case. Acta Anaesthesiol Scand 1981; 25: 115–16[ISI][Medline]
40 Eggert SM, Eggers KA. Subarachnoid haemorrhage following spinal anaesthesia in an obstetric patient. Br J Anaesth 2001; 86: 442–4 [Abstract/Free Full Text]
41 Eldor J. Opiate treatment of post-dural puncture headache. Acta Anaesthesiol Scand 1995; 39: 1140[ISI][Medline]
42 Eldor J, Guedj P, Cotev S. Epidural morphine injections for the treatment of postspinal headache. Can J Anaesth 1990; 37: 710–11[ISI][Medline]
43 Fink BR, Walker S. Orientation of fibers in human dorsal lumbar dura mater in relation to lumbar puncture. Anesth Analg 1989; 69: 768–72[Abstract]
44 Flaatten H, Rodt S, Rosland J, Vamnes J. Postoperative headache in young patients after spinal anaesthesia. Anaesthesia 1987; 42: 202–5[ISI][Medline]
45 Flaatten H, Rodt SA, Vamnes J, Rosland J, Wisborg T, Koller ME. Postdural puncture headache. A comparison between 26- and 29-gauge needles in young patients. Anaesthesia 1989; 44: 147–9[ISI][Medline]
46 Frumin MJ. Spinal anaesthesia using a 32-gauge needle. Anesthesiology 1969; 30: 599–60[ISI][Medline]
47 Geurts JW, Haanschoten MC, van Wijk RM, Kraak H, Besse TC. Post-dural puncture headache in young patients. A comparative study between the use of 0.52 mm (25-gauge) and 0.33 mm (29-gauge) spinal needles. Acta Anaesthesiol Scand 1990; 34: 350–3[ISI][Medline]
48 Gil F, Garcia-Aguado R, Barcia JA, et al. The effect of fibrin glue patch in an in vitro model of postdural puncture leakage. Anesth Analg 1998; 87: 1125–8[Abstract]
49 Gogarten W, Van Aken H. A century of regional analgesia in obstetrics. Anesth Analg 2000; 91: 773–5 [Free Full Text]
50 Gorelick PB, Zych D. James Leonard Corning and the early history of spinal puncture. Neurology 1987; 37: 672–4[ISI][Medline]
51 Gormley JB. Treatment of post-spinal headache. Anesthesiology 1960; 21: 565–6
52 Grant R, Condon B, Hart I, Teasdale GM. Changes in intracranial CSF volume after lumbar puncture and their relationship to post-LP headache. J Neurol Neurosurg Psychiatry 1991; 54: 440–2[Abstract]
53 Greene HM. Lumbar puncture and the prevention of post puncture headache. JAMA 1926; 86: 391–2
54 Gutterman P, Bezier HS. Prophylaxis of postmyelogram headaches. J Neurosurg 1978; 49: 869–71[ISI][Medline]
55 Handler CE, Smith FR, Perkin GD, Rose FC. Posture and lumbar puncture headache: a controlled trial in 50 patients. J R Soc Med 1982; 75: 404–7[Abstract]
56 Hannerz J, Ericson K, Bro Skejo HP. MR imaging with gadolinium in patients with and without post-lumbar puncture headache. Acta Radiol 1999; 40: 135–41[ISI][Medline]
57 Hansen PE, Hansen JH. DDAVP, a synthetic analogue of vasopressin, in prevention of headache after lumbar puncture and lumbar pneumoencephalograp hy. Acta Neurol Scand 1979; 60: 183–8[ISI][Medline]
58 Harrington H, Tyler HR, Welch K. Surgical treatment of post-lumbar puncture dural CSF leak causing chronic headache. Case report. J Neurosurg 1982; 57: 703–7[ISI][Medline]
59 Hart JR, Whitacre RG. Pencil point needle in the prevention of post-spinal headache. JAMA 1951; 147: 657–8[ISI][Medline]
60 Hebl JR, Horlocker TT, Chantigian RC, Schroeder DR. Epidural anesthesia and analgesia are not impaired after dural puncture with or without epidural blood patch. Anesth Analg 1999; 89: 390–4 [Abstract/Free Full Text]
61 Hodgson C, Roitberg-Henry A. The use of sumatriptan in the treatment of postdural puncture headache. Anaesthesia 1997; 52: 808[ISI][Medline]
62 Holder HG. Reactions after spinal anaesthesia. JAMA 1944; 124: 56–7
63 Holst D, Mollmann M, Ebel C, Hausman R, Wendt M. In vitro investigation of cerebrospinal fluid leakage after dural puncture with various spinal needles. Anesth Analg 1998; 87: 1331–5[Abstract]
64 Hoskin MF. Spinal anaesthesia—the current trend towards narrow gauge atraumatic (pencil point) needles. Case reports and review. Anaesth Intens Care 1998; 26: 96–106[ISI][Medline]
65 Jack TM. Post-partum intracranial subdural haematoma: a possible complication of epidural analgesia. Anaesthesia 1979; 34: 176–80[ISI][Medline]
66 Jarvis AP, Greenawalt JW, Fagraeus L. Intravenous caffeine for postdural puncture headache. Anesth Analg 1986; 65: 316–17[ISI][Medline]
67 Jokinen MJ, Pitkanen MT, Lehtonen E, Rosenberg PH. Deformed spinal needle tips and associated dural perforations examined by scanning electron microscopy. Acta Anaesthesiol Scand 1996; 40: 687–90[ISI][Medline]
68 Jones RJ. The role of recumbancy in the prevention and treatment of postspinal headache. Anesth Analg 1974; 53: 788–95[ISI][Medline]
69 Kang SB, Goodnough DE, Lee YK, et al. Comparison of 26- and 27-G needles for spinal anesthesia for ambulatory surgery patients. Anesthesiology 1992; 76: 734–8[CrossRef][ISI][Medline]
70 Keener EB. An experimental study of reactions of the dura mater to wounding and loss of substance. J Neurosurg 1959; 16: 424–47[ISI][Medline]
71 Khan KJ, Stride PC, Cooper GM. Does a bloody tap prevent postdural puncture headache? Anaesthesia 1993; 48: 628–9[ISI][Medline]
72 Klepstad P. Relief of postural post-dural puncture headache by an epidural blood patch 12 months after dural puncture. Acta Anaesthesiol Scand 1999; 43: 964–6[CrossRef][ISI]