Detection and management of epidural haematomas related to
anaesthesia in the UK: a national survey of current practice
, S. Bird
, J. J. Nightingale
* and N. White
Department of Anaesthesia, Basingstoke and North Hampshire Foundation Trust, Aldermaston Road,
Department of Anaesthesia, Queen Alexandra Hospital, Cosham, Portsmouth
Department of Anaesthesia, Royal Bournemouth Hospital, Castle Lane East, Bournemouth
*Corresponding author. E-mail: firstname.lastname@example.org
Background. Epidural haematoma is a rare, but potentially disastrous complication of epidural
analgesia. Favourable neurological outcome depends upon early recognition and surgical
decompression; therefore, the management of epidural analgesia should include a systematic
approach to recognition of the signs of epidural haematoma.
Methods. We conducted a national postal survey of the policies and protocols used by acute
pain services for investigating clinical signs suggestive of epidural haematoma, and the availability of
urgent MRI scans. This was a repeat of a survey that was carried out in 2001, but not published.
Results. The response rate was 84%. Of the acute pain services that responded, 99% have a
written protocol for running epidural infusions, 91% include regular assessment of sensory and
motor function, and 55% have a written protocol for the investigation of abnormal motor block.
On-site 24 h access to MRI scanning facilities was available to 57%, 33% have arrangements with
another hospital, and 10% do not have 24 h access to MRI. Thirty per cent of respondents knew
of an epidural haematoma related to epidural analgesia in their hospital, one-third of which were
not diagnosed and treated within 24 h.
Conclusions. Improvements in monitoring have occurred over the last 5 yr, but observations of
neurological function are not routine in all units, and are not continued after removal of the epi-
dural catheter in the majority. The authors suggest that acute pain services should be responsible
for protocols for the investigation and treatment of epidural haematomas.
Br J Anaesth 2008; 101: 400–4
Keywords: anaesthetic techniques, epidural; analgesia, postoperative; analgesic techniques,
epidural; complications, haematoma; complications, neurological
Accepted for publication: May 4, 2008
Epidural infusions are used routinely for analgesia after
operation and during labour. The recently published
national census of central neuraxial block in the UK, which
reported the snapshot phase of the Third National Audit
Project of the Royal College of Anaesthetists,
found that a
bined spinal and epidurals were performed for postopera-
tive analgesia in adults and children over a 2 week period.
Applying the multiplier of 25 used by the census, the
authors suggest that approximately 112 400 epidurals may
be performed annually for postoperative analgesia in the
UK. The survey does not distinguish between epidural
catheter insertion and ‘single-shot’ epidurals, so the rate of
catheter insertion may be lower, but it is reasonable to
assume that the majority of these procedures involve inser-
tion of an epidural catheter. This correlates with our esti-
mate obtained by extrapolating from our acute pain service
audit data. Approximately 900 epidural catheters are sited
annually for postoperative analgesia in Portsmouth, which
serves a population of approximately 580 000—roughly
1% of the national population, so assuming other insti-
tutions have a similar epidural rate to our own, the number
This article is accompanied by Editorial I.
British Journal of Anaesthesia 101 (3): 400–4 (2008)
Advance Access publication June 13, 2008
analgesia in the UK would be of the order of 90 000.
The quoted incidence of epidural haematomas is around
but this is likely to be an underestimate as it
suggest a likely incidence of one epidural haematoma
every 2 yr related to epidurals used for postoperative
analgesia in the UK. Although rare, the consequences of
an epidural haematoma can be devastating, especially if
not detected and treated rapidly.
Seven years ago, an epidural haematoma occurred in our
hospital. Both detection and treatment were delayed. In
response to this, we revised the protocol for investigation
of abnormal motor block and, as MRI was not routinely
available outside normal working hours in our hospital,
made arrangements with the regional neurosurgery unit for
MRI to be performed when indicated by our acute pain
service protocol. We subsequently undertook, but did not
publish, a national postal survey to assess how and where
patients with epidural infusions were monitored in other
hospitals, and what arrangements were in place for the
investigation and management of epidural haematomas.
The results of this survey showed that regular checks of
sensory and motor function did not occur in all hospitals
and that fewer than one-third of acute pain services contin-
ued checks after the epidural had been removed. Only
43% had access to MRI scanning in their hospital. Many
of those without direct MRI access used MRI scanners in
non-neurosurgical units, increasing treatment delay if the
results were positive. Thus, 6 yr later, we have repeated
the survey to elicit current practice and to determine
whether practice in this area had changed.
We obtained a list of all anaesthetic departments in the
the addresses of the 301 departments registered as having
College Tutors. We sent a numbered questionnaire
(Appendix 1) to each of these departments, addressed to a
member of the acute pain service. A covering letter
explained that data collected would be made anonymous
and be non-attributable. After 12 weeks, we re-sent the
questionnaire to non-responders. The results were collated
and compared with those from 2001.
254 replies were returned—a response rate of 84%. Ten hos-
pitals which had no acute pain team returned uncompleted
questionnaires. All but two hospitals ran postoperative
epidurals. Thus, completed replies were received from 242
hospitals where postoperative epidural infusions were in use,
and this ﬁgure is used as the denominator for calculating
percentages, which are rounded to the nearest integer.
Epidural infusions are managed on normal wards in 222
units (92%), and the remaining 20 units (8%) run them
only on a high dependency facility. These results show a
change in practice from 2001 when only 80% of units
managed epidurals on surgical wards. Written protocols for
running postoperative epidural infusions were in place in
239 units (99%), compared with 95% in 2001 (Table 1).
Two hundred and twenty units (91%) make regular assess-
ment of sensory level and motor function (84% in 2001),
with the remaining 20 not making assessments (Table 1).
In 189 units (78%), these observations are made at least 4
hourly (six respondents did not answer this question). One
hundred and seven units (44%) continue to monitor
sensory and motor function after epidural catheters are
removed (29% in 2001), and 30 units (12%) monitor for
more than 12 h after removal.
In 2001, only 31 units (13%) had a written protocol for
the investigation of a suspected epidural haematoma. In
2007, we asked more speciﬁcally about the existence of a
protocol for the investigation of abnormal motor block,
and 129 (53%) conﬁrmed such a protocol was in place.
When an epidural haematoma is suspected, a consultant
anaesthetist is solely (49%) or jointly (34%) responsible
for instigating investigation in 202 units, leaving 40 units
(17%) in which responsibility is not taken by a consultant
Access to MRI scans
Six years ago, 43% of units had 24 h access to an MRI
scanner, compared with 136 (57%) now, but 100 do not,
and six did not answer this question (Table 2). Of the 100
units which do not have in-house access to a scanner, 81
have access to an MRI in another hospital and 19 do not
have 24 h access. One hundred and twenty-six units (52%)
have a speciﬁc agreement with their radiologists to allow
24 h access to MRI scanning for suspected haematomas
Acute pain service protocols
Do you have a written
protocol for running
Does this include regular
assessment of sensory level
and motor function?
Are observations made at
least 4 hourly?
Do observations continue
after the epidural is
than 12 h?
Detection and management of epidural haematomas related to anaesthesia
units (19%) have agreed an investigation and treatment
protocol with their local neurosurgical or spinal unit, 187
(77%) do not, and 11 did not answer this question.
Seventy-two (30%) of the respondents were aware of an
epidural haematoma that had occurred in their hospital at
some time. The ﬁgure 6 yr previously was 32 (13%). This
suggests that our estimate of one epidural haematoma per
2 yr occurring in the UK may indeed be an underestimate.
In 48 of 72 cases of epidural haematoma, the diagnosis,
investigation, and treatment were achieved in 24 h, but in
24 cases (33%), this was not achieved. Reasons for delay
reported in the current survey were: delay in picking up
the clinical signs (20 cases), lack of MRI availability (two
cases), and delays in transfer to other units (two cases). In
comparison, in 2001, 10 out of 32 cases (31%) were not
managed within 24 h.
Epidural haematomas after the use of epidural analgesia are
mercifully rare. However, in our study, 72 respondents were
aware of an epidural haematoma having occurred in their
unit, an increase of 40 over the last 6 yr. The forthcoming
publication of the second stage of the Royal College of
Anaesthetists Third National Audit project may provide more
information about the incidence of epidural haematomas.
Classically, epidural haematomas cause radicular pain,
motor impairment, sensory loss, and urinary retention.
Patients with epidural infusions are usually catheterized
and have some sensory deﬁcit, so these clinical signs may
not be helpful in detecting problems. When related to
epidural anaesthesia, not all haematomas are painful. In a
review of 61 epidural haematomas related to central neur-
axial block, pain was the presenting complaint in only
38% of cases.
The most reliable sign of a developing hae-
opment of motor block, and this should therefore be
checked for regularly. The literature also suggests that
motor block is the most sensitive prognostic indicator. The
Frankel scale for assessing spinal injury and its subsequent
modiﬁcation by the American Spinal Injury Association
uses motor function as the principle variable assessed
(Appendix 2). This is the scoring system used in the two
largest studies relating outcome from epidural haematomas
to the severity of neurological deﬁcit.
checks of motor and sensory block are not made, although
this has improved a little compared with 6 yr ago. The
authors agree with Christie and McCabe
that it is vital
with epidural analgesia.
Epidural haematoma related to an epidural catheter may
reported that 50% of haematomas
relating to epidural catheters occurred after their removal.
There have been several case reports of haematomas
occurring more than 12 h after catheter removal.
of concern that only 44% of units continue epidural block
assessments after removal of the catheter, and only 12%
of units continue these assessments for more than 12
h. The authors believe that observations should be contin-
ued for 24 h after removal of the catheter.
The deﬁnitive treatment of an epidural haematoma is sur-
gical decompression by laminectomy. The factors that deter-
mine outcome are the severity of the neurological deﬁcit at
presentation and the time from presentation to surgery.
surgery is carried out within 12 h of the onset of symptoms,
more than 24 h after the presentation of symptoms, recovery
rates drop to about 10%.
This demonstrates the importance
while the catheter is in place and after it is removed. We feel
that the referral path for patients with signs suggestive of
epidural haematoma, particularly motor block, should be
clearly delineated, preferably involving a consultant anaes-
thetist, in order to avoid delays in the instigation of appro-
priate investigations and treatment.
The investigation of choice for a suspected epidural
haematoma is an MRI scan.
Our study showed that 56%
increase of 25% over 6 yr. Twenty-four units did not have
any access to MRI scanning. Most of the units without
MRI scanners have access to a scanner in another hospital.
However, if this is not the hospital that can offer decom-
pression, then a positive result will need a second transfer
to a neurosurgical or spinal unit. As time between diagno-
sis and treatment is critical, introducing another transfer
Facilities for investigating and treating haematoma
Do you have access to
emergency MRI scanning
24 h a day in your
If not do you have 24 h
access in another hospital?
Do you have an agreement
with your radiologists to
provide urgent MRI scans
for suspected epidural
haematomas 24 h a day?
Do you have an
protocol agreed with your
local spinal or
would suggest that if a hospital does not have access to an
MRI scanner, patients in whom an epidural haematoma is
suspected would most appropriately be sent for MRI scan
to a unit which has the capacity to operate.
Epidural haematomas are not common and early detec-
tion and treatment can make a profound difference to
outcome. Therefore, the authors believe that it is necessary
to have protocols for the management of suspected cases,
covering assessment of motor and sensory function, access
to MRI scanning, and referral to a neurosurgical unit. To
make the comparison with malignant hyperpyrexia, which
is similarly rare (1:40 000 patients),
there are widely
same practice should be used for epidural haematomas.
After reviewing the literature and examining practice, we
propose that a protocol for the diagnosis, investigation,
and management of epidural haematomas should include
the following elements:
(i) Patients with epidural infusions running should have
observations that include assessment of motor block
made at least every 4 h.
(ii) These observations should continue for at least 24 h
after removal of the epidural catheter.
(iii) There should be a designated person responsible for
investigating signs suggestive of epidural haematoma.
(iv) If signiﬁcant deterioration in motor function occurs
in the absence of a recent bolus dose of local anaes-
thetic being administered, the designated person
should be contacted immediately.
(v) If motor block is attributed to a recent bolus dose of
epidural drugs, reassessment should occur within 2 h.
(vi) If an epidural infusion is running, it should be
turned off, alternative analgesia instigated as necess-
ary, and a reassessment of the patient’s motor func-
tion should be made after a deﬁned interval. The
motor block would be expected to resolve if due to
overdose or catheter migration. If motor power does
not improve, remediable causes, including epidural
haematoma or abscess, must be excluded.
(vii) Once an epidural haematoma is suspected, an MRI
scan should be organized immediately, as this is a
(viii) If MRI scanning is not available in the local hospi-
tal or there will be a delay, then the patient should
be referred to a neurosurgical unit to be scanned. It
may be appropriate to arrange a protocol with local
neurosurgical units to minimize delays in investi-
gation and treatment.
The authors would like to thank all those who completed and returned
the data collection forms.
Trust Anaesthetic Department.
Appendix 1: Data collection form
Survey of management and investigation of spinal –
epidura haematomas secondary to epidural analgesia
1. Do you use epidurals for peri and postoperative
analgesia in your hospital?
A Yes A No
2. Do you run epidural infusions postoperatively?
3. If yes, where are the patients managed?
A ITU A HDU A Ward
4. Do you have an aute pain service?
5. Do you have a written protocol for running
postoperative epidural infusions?
A Yes A No
6. Does this include regular assessent of sensory level
and motor function?
7. How frequently are these obsevations made?
A ,1H A 1 – 2H A 2 – 4H A 2 – 4H A .4H
8. Do they continue after the epidural is removed?
9. If yes to question 8, for how long?
A 0 – 4H A 4 – 8H A 8 – 12H A .12H
10. Do you have a written protocol for the investigation of
abnormal motor block?
11. Who takes responsibility for investigation of suspected
A Surgical team A Pain nurse A Anaesthetic trainee
A Anaesthetic consultant
12. Do you have access to emergency MRI scanning 24
hours per day in your hospital?
13. If the answer to question 12 is no, do you have access
to scaning 24 hours per day in another hospital?
14. If the answer to question 13 is yes, is this hospital
your local spinal or neurosurgery unit?
15. If you have answered yes to question 12 or 13, have
your radiologists agreed to provide urgent MRIs for
suspected epidural haematomas 24 hours per day?
16. Do you have an investigation and treatment protocol
agreed with your local spinal surgery of neurosurgical
epidural anesthesia or analgesia in your hospital?
18. If yes, was it diagnosed, investigated and treated
within 24 hours?
19. If not where was the worst delay
A Clinical diagnosis
A Getting an MRI scan
A Waiting for transfer to spinal injury unit
Appendix 2: Frankel scale for spinal injury
American Spinal Injury Association (ASIA) impairment
A, complete. No sensory or motor function is preserved in
the sacral segments S4 – S5.
preserved below the neurological level and includes the
C, incomplete. Motor function is preserved below the
neurological level and more than half of key muscles
below the neurological level have a muscle grade less
D, incomplete. Motor function is preserved below the
neurological level and at least half of key muscles
below the neurological level have a muscle grade
greater than or equal to three.
E, normal. Sensory and motor function are normal.
1 Cook T, Mihai R, Wildsmith JAW. A national census of central
neuraxial block in the UK: results of the snapshot phase of the
Anaesthetists. Anaesthesia 2008; 63: 143 – 6
2 Wulf H. Epidural anaesthesia and spinal haematoma. Can J Anaesth
1996; 43: 1260 – 71
3 Vandermeulen EP, Van Aken H, Vermelyn J. Anticoagulants
and spinal – epidural analgesia. Anesth Analg 1994; 79: 1165 – 77
4 Lawton MT, Porter RW, Heiserman EJ, et al. Surgical management
of spinal epidural haematoma: relationship between surgical
timing and neurological outcome. J Neurosurg 1995; 83: 1 – 7
5 Groen RJM, Van Alphen HAM. Operative treatment of spon-
taneous spinal epidural haematomas: a study of the factors
determining postoperative outcome. Neurosurgery 1996; 39:
494 – 508
6 Christie IW, McCabe S. Major complications of epidural analgesia
after surgery: results of a six-year survey. Anaesthesia 2007; 62:
335 – 41
epidural catheter. Reg Anaesth 1991; 14: 88 – 91
8 Ousmane ML, Fleyfel M, Vallet B. Epidural hematoma after catheter
removal. Anesth Analg 2000; 90: 1250
9 Sandhu H, Morley-Forster P, Spadafora S. Epidural hematoma
following epidural analgesia in a patient receiving unfractionated
heparin for thromboprophylaxis. Reg Anesth Pain Med 2000; 25:
72 – 5
removal of an epidural catheter in a patient receiving high-dose
enoxaparin. Br J Anaesth 1999; 82: 288 – 90
11 Larsson EM, Holtas S, Cronqvist S. Emergency magnetic resonance
examination of patients with spinal cord symptoms. Acta Radiol
1988; 29: 69– 75
12 Abraham RB, Adnet P, Glauber V, Perel A. Malignant hyperthermia.
Postgrad Med J 1998; 74: 11– 7