From the Society for Vascular Surgery
Complications of spinal fluid drainage in
thoracoabdominal aortic aneurysm repair: A report
of 486 patients treated from 1987 to 2008
Martha M. Wynn, MD, Matthew W. Mell, MD, Girma Tefera, MD, John R. Hoch, MD, and
Charles W. Acher, MD, Madison, Wis
Objective: Spinal fluid drainage reduces paraplegia risk in thoracic (TAA) and thoracoabdominal (TAAA) aortic aneurysm
repair. There has not been a comprehensive study of the risks of spinal fluid drainage and how these risks can be reduced.
Here we report complications of spinal fluid drainage in patients undergoing TAA/TAAA repair.
Methods: The study comprised 648 patients who had TAA or TAAA repair from 1987 to 2008. Spinal drains were used
in 486 patients. Spinal fluid pressure was measured continuously, except when draining fluid, and was reduced to <6 mm
Hg during thoracic aortic occlusion and reperfusion. After surgery, spinal fluid pressure was kept <10 mm Hg until
patients were awake with normal leg lift. Drains were removed 48 hours after surgery. Spinal and head computed
tomography (CT) scans were performed in patients with bloody spinal fluid or neurologic deficit. We studied the
incidence of headache treated with epidural blood patch, infection, bloody spinal fluid, intracranial and spinal bleeding
on CT, as well as the clinical consequences.
Results: Twenty-four patients (5%) had bloody spinal fluid. CT exams showed seven had no evidence of intracranial
hemorrhage, 14 (2.9%) had intracranial blood without neurologic deficit, and three with intracranial bleeding and
cerebral atrophy had neurologic deficits (1 died, 1 had permanent hemiparesis, and 1 with transient ataxia recovered
fully). Two patients without bloody spinal fluid or neurologic deficit after surgery presented with neurologic deficits 5
days postoperatively and died from acute on chronic subdural hematoma. Neurologic deficits occurred after spinal fluid
drainage in 5 of 482 patients (1%), and 3 died. The mortality from spinal fluid drainage complications was 0.6% (3 of
482). By univariate and multivariate analysis, larger volume of spinal fluid drainage (mean, 178 mL vs 124 mL, P <
.0001) and higher central venous pressure before thoracic aortic occlusion (mean, 16 mm Hg vs 13 mm Hg, P < .0012)
correlated with bloody spinal fluid.
Conclusion: Strategies that reduce the volume of spinal fluid drainage but still control spinal fluid pressure are helpful in
reducing serious complications. Patients with cerebral atrophy are at increased risk for complications of spinal fluid
drainage. (J Vasc Surg 2009;49:29-35.)
Spinal fluid drainage has become a standard interven-
tion to reduce paraplegia risk in thoracic (TAA) and thora-
coabdominal (TAAA) aortic aneurysm repair.1-4Complica-
tions associated with diagnostic lumbar puncture,5neuraxial
anesthesia,6and spinal fluid drainage7,8have been reported.
Although spinal fluid drainage has been used in TAAA
repair for 20 years,9there has not been a comprehensive
study of the associated risks. Here we report a retrospective
analysis of the complications of spinal fluid drainage in 486
A concurrently maintained, institutionally approved
database was used to study 648 patients who had TAA or
TAAA repair from 1987 to 2008. Surgery in all patients
who had open repair with spinal fluid drainage was done
without assisted circulation or heparinization. Since 2005,
thoracic aneurysms have been treated with endografts.
clotting time of 200 to 250 seconds.
Spinal drains were placed preoperatively by anesthesi-
ologists in 486 patients. Drains were not placed in acute
disease, or patients with abnormal results on coagulation
studies. Elective patients stopped clopidogrel 7 days before
surgery, but continued aspirin. If blood could be aspirated
from the needle during drain insertion, elective surgery was
postponed. From 1987 to 2000, a small (19-gauge) epi-
dural catheter was placed using anatomic landmarks, with
needle placement at L3-L4 or L2-L3 and the catheter
advanced 10 cm after entering the dura. Because of an
unacceptable number of drain failures and difficult drain
insertions, after 2000 a larger (16-gauge) Silicone drain
(Medtronic EMD lumbar drain, Goleta, Calif) was placed
under fluoroscopy, with needle insertion at L3-L4 or
L2-L3 and the catheter tip positioned at T9-T10.
Spinal fluid pressure (SFP) was measured immediately
after drain insertion, continuously during surgery except
when fluid was drained to gravity in 5-mL increments, and
for 48 hours after surgery. In open repairs spinal fluid was
From the University of Wisconsin School of Medicine and Public Health.
Competition of interest: none.
Presented at the Society for Vascular Surgery meeting, San Diego, Calif,
June 5-8, 2008.
Correspondence: Dr Martha Marie Wynn, Anesthesiology, University of
Wisconsin School of Medicine & Public Health, 600 Highland Ave,
Madison, WI 53705 (e-mail: email@example.com.)
Copyright © 2009 Published by Elsevier Inc. on behalf of The Society for
drained to achieve a SFP of ?6 mm Hg during surgery.
Beginning in 2003, spinal fluid was drained to achieve a
SFP of ?6 mm Hg only during thoracic aortic occlusion
periods of the surgery by draining the smallest volume of
spinal fluid possible. In thoracic endograft procedures, SFP
was reduced to ?10 mm Hg before device deployment.
The total volume of spinal fluid drained during surgery
was measured. Drain failure was defined as being unable to
drain enough fluid to control SFP during thoracic aortic
occlusion. After surgery, spinal fluid was drained to keep
the SFP ?10 mm Hg only until patients were awake with
normal leg lift. After normal leg lift was observed, SFP was
monitored, and fluid was not drained unless weakness
occurred. Drains were removed 48 hours after open repairs
and 24 hours after endograft procedures.
Drainage was stopped if the fluid became bloody. Spi-
nal and head computed tomography (CT) scans were per-
formed in patients with bloody spinal fluid and in all
patients with abnormal neurologic signs. Postoperative
postural headache was treated with epidural blood patch.
Patients were followed up for 12 months after surgery.
We studied the incidence of epidural blood patch, postop-
erative neurologic deficit, infection, bloody spinal fluid,
presence of spinal and intracranial bleeding on CT, and the
resulting clinical consequences. The Fisher exact test and
The mean age was 67 years, and 54% of patients were
men. There were 72 patients (15%) with TAA, and 25
(34%) were treated with endografts. There were 414 pa-
tients with TAAAs, consisting of Crawford category type 1
in 73 (18%), type 2 in 139 (34%), type 3 in 89 (21%), and
type 4 in 113 (27%). There were 135 acute patients (28%),
with rupture, acute dissection, aortitis, or trauma. The
30-day mortality was 3.8% in elective patients and 13.3% in
those presenting acutely. A total of 311patients (64%) had
a small drain, and 175 (36%) had a large drain. Four acute
patients died intraoperatively and could not be analyzed for
drain complications. Twenty-three patients (4.8%) were
paralyzed, and six of these had delayed paralysis.
There was no difference in mean volume of fluid
drained in small drain (127 mL) and large drain (128 mL)
patients, or in mean volume of fluid drained in patients
treated from 1987 to 2003 (130 mL) and those treated
from 2003 to 2008 (120 mL; P ? .074). Drain failure
occurred in 24 of 308 patients (7.8%) with small drains and
in three of 174 patients (1.7%) with large drains placed
using fluoroscopy; this difference was significant (P ?
.0054). Since using fluoroscopy, we were technically un-
able to place a drain in only one patient and did not aspirate
blood during drain placement in any patient. Two small-
drain patients (0.6%) and four large-drain patients (2.3%)
had postural headaches that were treated with epidural
blood patch; this difference was not significant. There were
no spinal drain infections.
they underwent spinal and head CT scans. No patient had
spinal hematoma on CT. Seven of the 24 had no CT
evidence of an intracranial hemorrhage. Head CT scans
showed intracranial bleeding in 17 patients (3.5%): 10
patients had small, five had moderate, and two had large
Six patients had small or moderate intraparenchymal bleed-
ing only, without an accompanying subarachnoid or sub-
dural hematoma. No epidural hematomas were found.
Of the 17 patients with intracranial blood on CT, 14
patients without deficits, small acute hemorrhages were
superimposed on an existing undiagnosed intracranial le-
sion: one patient had new hemorrhage at the site of an old
subdural hematoma, one had bleeding near a small menin-
Table I. Location of hemorrhage on computed tomography scan in patients with bloody spinal fluid
L frontalL frontal
R parietal, R/L sulci
Ataxia that resolved
TentoriumR lateral fissure
R frontal/parietalR frontal/parietalL hemiplegia, ambulatory
L, left; Lrg, Large; M, moderate; R, right; S, small.
JOURNAL OF VASCULAR SURGERY
30 Wynn et al
gioma, and one had new hemorrhage at the site of an old
Neurologic deficits were present in three patients with
intracranial bleeding on CT: one died from brain hernia-
tion caused by mass effect, one sustained permanent left
hemiparesis but was ambulatory; and one recovered fully
from transient ataxia. Both the patient who died and the
patient with hemiparesis had large right-sided cerebral sub-
dural hematomas. The patient with transient ataxia had a
moderate-sized left cerebellar intraparenchymal hemor-
rhage (Table I). All three patients had pre-existing, but
unknown, cerebral atrophy with brain volume loss.
Two patients who did not have bloody spinal fluid or
neurologic deficit after surgery developed neurologic defi-
cits on postoperative day 5 (3 days after spinal drain re-
moval) after receiving enoxaparin for deep venous throm-
bosis and heparin for pulmonary embolus. Both had acute
cerebral subdural hematomas at the site of an undiagnosed
chronic subdural hematoma on CT and died despite neu-
rosurgical evacuation of the hematoma.
Neurologic deficits occurred in five of 482 patients
(1%) after spinal fluid drainage, and three of these patients
died. The mortality from spinal drain complications was
0.6% (3 of 482). By univariate and multivariate analysis,
higher central venous pressure before aortic occlusion
(mean, 16 vs 13 mm Hg, P ? .0012) and volume of spinal
fluid drained intraoperatively (mean, 178 vs 124 mL, P ?
.0001) correlated with bloody spinal fluid (Tables II, III
and IV). Age, gender, drain type, SFP, acuity, blood re-
placement, Crawford aneurysm type, renal ischemia time,
preoperative renal function, arterial blood pressure, and
other hemodynamic variables were not significant. No pa-
tient presented with a new neurologic deficit after dis-
The choroid plexus and walls of the ventricles produce
cerebrospinal fluid at a rate that varies in a diurnal rhythm
from 0.2 to 0.7 mL/min, producing a total of 400 to 600
mL of fluid each 24 hours.10Cerebrospinal fluid circulates
continuously from the brain to the spinal canal and back to
the brain, where it is reabsorbed in the venous sinuses. The
total volume of circulating fluid is about 140 mL in an
adult, and the ventricles hold an additional 25 mL.10
Changes in spinal fluid volume and pressure affect
intracranial mechanics and can produce deleterious clinical
intracranial hypotension. Headache is a well-known result
of procedures that reduce spinal fluid volume and pressure,
such as lumbar puncture, spinal anesthesia, myelography,
and ventricular shunts.11The mechanism of headache is
thought to be tension on the sensory receptors of the dural
sinuses.12,13Headache requiring treatment with epidural
blood patch was a complication of spinal fluid drainage
during TAAA repair. CT and magnetic resonance scans
done after spinal fluid removal in diagnostic lumbar punc-
ture show caudal displacement of the brain, with cerebellar
tonsillar sagging, a decrease in ventricular size, and dilated
veins.14These findings correlate with decreased SFP.15
Venous engorgement after lumbar puncture may cushion
the displacement of the brain caused by spinal fluid re-
Intracranial hypotension can cause acute intracranial
bleeding. This complication has been documented after
diagnostic lumbar puncture,11spinal anesthesia,6ventricu-
loperitoneal shunt placement, myelography,13spinal fluid
drainage during neurosurgery or thoracic aortic surgery,7,8
and dural leak after spine surgery.13Some series have
reported subdural hematomas in 10% of cases of intracra-
nial hypotension, regardless of the cause.11
Table II. Univariate Analysis for Intracranial Bleeding
Age, years 67.76
SF drained, mL
SFP-1, mm Hg.613
SFP-2, mm Hg.741
SFP change, mm Hg .575
CVP-1, mm Hg.0012a
CVP-2, mm Hg.0034
CVP-3, mm Hg .0178
MAP-1, mm Hg .2363
MAP-2, mm Hg98.33
MAP-3, mm Hg .5291
Blood replacement, L.0735
Pre-op creatinine, mg/dL.5527
1, Preclamp; 2, during clamp; 3, unclamped; CVP, central venous pressure;
MAP, mean arterial pressure; PAP, pulmonary artery pressure; SFP, spinal
aVolume of spinal fluid drainage and higher CVP were significant for
bleeding on univariate analysis.
Table III. Univariate analysis for intracranial bleedinga
Variable BleedingTotal No.%
aAge, gender, and acuity were not significant.
JOURNAL OF VASCULAR SURGERY
Volume 49, Number 1
Wynn et al 31
Loss of spinal fluid reduces intracranial pressure, lead-
brain displacement from lower intracranial pressure creates
tension on enlarged venous sinuses that predisposes to
venous tears.12Intracranial hypotension can stretch and
rupture large cortical veins crossing the subdural space.17
Reflex vasodilatation in response to pressure on the dura,
veins, and dural sinuses by the caudally displaced brain may
increase the risk of subdural bleeding.18
Bleeding can occur in the posterior fossa under the
tentorium cerebellum as the cerebellum sags onto the
foramen magnum, above the tentorium as the cerebrum
sags in a caudal fashion onto the cerebellum, and in the
temporal and parietal lobes as the cerebrum sags onto the
cranium. If the sag is slow, there may be no bleeding at all;
but if the brain deforms rapidly, the resulting tension may
tear subdural veins, producing bleeding. Intracranial hypo-
tension can also produce subarachnoid hemorrhage.13Pa-
tients with a history of head trauma, chronic subdural
hematoma, cerebral atrophy, cranial vault abnormalities,
arteriovenous malformations, abnormalities of coagula-
tion, or cerebral aneurysms may be at increased risk for
hemorrhagic complications of intracranial hypotension.12
SFP normally approximates central venous pressure,19
and spinal cord perfusion pressure is the difference between
mean arterial pressure and SFP. Thoracic aortic clamping
produces an acute rise in both central venous pressure and
SFP. This increase in SFP decreases spinal cord blood
flow.20,21Animal experiments using thoracic aortic occlu-
sion models show that draining spinal fluid during aortic
occlusion reduces SFP, improves spinal cord perfusion, and
reduces paralysis.22,23Spinal fluid drainage has also been
shown to reduce paralysis in randomized trials of patients
having TAAA repair.24,25Although there are proven ben-
efits to spinal fluid drainage, less is known about the risks.
Whether neuraxial bleeding occurs from spinal fluid
drainage during TAAA repair and whether it results in
neurologic deficit probably depends on factors such as
traumatic drain placement, volume and rate of spinal fluid
drainage, extent of venous engorgement, size of the veins
that rupture, pre-existing intracranial pathology, and the
presence of coagulopathy. The most serious complication
of spinal fluid drainage was intracranial bleeding, and not
intraspinal, spinal subarachnoid, epidural or subdural
bleeding at the site of needle insertion or catheter place-
ment, although others have reported intraspinal bleed-
ing.26In our experience the appearance of blood in the
spinal fluid was a very sensitive indication of intracranial
bleeding, even in the absence of clinical neurologic signs.
Bloody spinal fluid occurred with both small and large
drains and correlated with draining a larger volume of fluid
during surgery. Higher central venous pressure before tho-
racic aortic occlusion also correlated with bloody spinal
fluid. Arterial hypertension did not.
Subdural27,28and intraparenchymal29intracranial he-
matoma are recognized complications of spinal fluid drain-
by Dardik et al,28draining a larger volume of fluid (690 vs
359 mL) in the perioperative period increased the risk that
a subdural hematoma would develop.28These totals in-
cluded significant drainage after surgery.
We do not believe that postoperative drainage is neces-
sary or desirable as long as patients have normal leg
strength. Because we let SFP return to baseline as soon as
fluid drainage. However, draining a larger volume of fluid
intraoperatively (178 vs 124 mL) was a significant risk
factor for intracranial bleeding. Whether the pressure
threshold for drainage should be 12, 10, 5, or 0 cm H2O or
target pressures somewhat arbitrarily because there is no
evidence to prove which pressure has better outcomes.
Because most of the patients in the Dardik et al study
with subdural hematomas were drained to a pressure of 5
cm H2O (3.8 mm Hg), they recommended draining only
to a pressure of 10 cm H2O (7.7 mm Hg).28However, our
patients all had spinal fluid drainage to ?6 mm Hg, and
lower SFP was not a risk factor in intracranial bleeding.
Dardik reported that only two of eight subdural hemato-
mas appeared in the first week, and two were diagnosed 1.5
and 5 months after surgery.28We did not have any late
subdural hematomas. All patients were monitored for 1
year after discharge, and no patient presented with neuro-
logic deficit or clinical signs of subdural hematoma ?5 days
Moderately small amounts of subdural, subarachnoid,
or intraparenchymal blood did not result in neurologic
deficit. Each of the three patients that presented with
immediate neurologic deficits had a large intracranial hem-
orrhage on CT. These patients also had unknown cerebral
atrophy, with loss of brain volume, that may have affected
the size of the initial hemorrhage or, in the case of subdural
hematoma, whether it became large enough to cause neu-
rologic deficit or death from intracranial mass effect. The
Table IV. Multivariate analysis
VariableProbability OR95% CI
Spinal fluid removed
1, Preclamp; 2, during clamp; CI, confidence interval; CVP, central venous pressure; OR, odds ratio.
CVP before aortic occlusion and volume of spinal fluid drained remained significant on multivariate analysis.
JOURNAL OF VASCULAR SURGERY
32 Wynn et al
two patients who did not have bloody spinal fluid or
neurologic deficit after surgery that developed delayed
neurologic deficits on postoperative day 5 after being anti-
coagulated both had an acute subdural cerebral hematoma
at the site of an undiagnosed old subdural hematoma on
CT. In these patients, 115 and 120 mL of spinal fluid was
removed without immediate complications, and we believe
anticoagulation in the presence of an old subdural hema-
toma also played a role in their hemorrhages.
Epidemiologic studies report the incidence of chronic
subdural hematoma as 7.35 to 13/100,000 population in
persons aged ?65 years and up to 58/100,000 in persons
?70 years.30,31These estimates were extrapolated from
neurosurgical procedures in elderly populations and may
actually underestimate the incidence of asymptomatic sub-
dural hematomas. Patients with pre-existing intracranial
disease that produces loss of brain volume, those with
hydrocephalus, and those with old subdural hematoma
may be at more risk of intracranial bleeding complications
from spinal fluid drainage during TAAA repair. Although
we believe patients with pre-existing intracranial abnormal-
ities are more likely to have serious neurologic deficits from
intracranial bleeding after spinal fluid drainage, it is not
always possible to identify them before surgery. However,
elective patients with abnormal neurologic findings, alco-
holism, and previous head injury or neurosurgery now
undergo preoperative head CT scanning.
It is important to note that although most patients with
bloody spinal fluid had CT evidence of intracranial hemor-
rhage, most patients with CT evidence of bleeding did not
have neurologic deficits. If spinal fluid became bloody, we
stopped draining fluid immediately and corrected any co-
agulopathy that could increase the size of the hematoma
and the likelihood it would produce neurologic deficit.
are important measures in reducing neurologic morbidity
and mortality from spinal fluid drainage complications.
Controlling SFP with the minimum volume of spinal
fluid drainage probably reduces intracranial bleeding com-
plications. The total amount of fluid drained depends in
part on the duration of surgery and aortic occlusion. How-
ever, timing spinal fluid drainage so that the SFP is not
reduced too early in surgery has allowed us to decrease the
volume of fluid drained and still control spinal fluid pres-
sure at critical times during surgery. We suspect that using
this strategy in the last 5 years has helped to reduce the
occurrence of large intracranial bleeds from spinal fluid
drainage that cause neurologic deficits; however, this dif-
ference in practice over time was not significant on analysis.
Using strategies to decrease the volume of spinal fluid
drained, but still control SFP, may reduce serious compli-
cations of spinal fluid drainage during TAAA repair. Pa-
tients with cerebral atrophy or unrecognized chronic sub-
dural hematoma may be at increased risk for serious
neurologic complications after spinal fluid drainage.
Conception and design: MW, GT, JH
Analysis and interpretation: MW, CA, MM
Data collection: MW, CA
Writing the article: MW, CA
Critical revision of the article: MW, CA
Final approval of the article: MW, CA, MM, GT, JH
Statistical analysis: CA
Obtained funding: Not applicable
Overall responsibility: MW
1. Coselli JS, Bozinovski J, LeMaire SA. Open surgical repair of 2286
thoracoabdominal aortic aneurysms. Ann Thorac Surg 2007;83:
2. Safi HJ, Miller CC, Huynh TT, Estrera A, Porat E, Winnerkvist A, et al.
Distal aortic perfusion and cerebrospinal fluid drainage for thoracoab-
dominal and descending thoracic aortic repair: ten years of organ
protection. Ann Surg 2003;238:372-80.
3. Jacobs MJ, Mess W, Mochtar B, Nijenhuis R, Stratius van Eps R,
Schurink G. The value of motor evoked potentials in reducing paralysis
during thoracoabdominal aneurysm repair. J Vasc Surg 2006;43:
Looking for the artery of Adamkiewicz: a quest to minimize paraplegia
after operations for aneurysms of the descending thoracic and thoraco-
abdominal aorta. J Thorac Cardiovasc Surg 1996;112:1202-13.
5. Ruff RL, Dougherty JH. Complications of lumbar puncture followed
by anticoagulation. Stroke 1981;12:879-81.
6. Horlocker TT, Wedel DJ, Benzon H, Brown DL, Enneking F, Heit J,
et al. Regional anesthesia in the anticoagulated patient: Defining the
risks (The Second ASRA Consensus Conference on Neuraxial Anesthe-
sia and Anticoagulation). Reg Anesth Pain Med 2003;28:172-97.
7. Weaver K, Wiseman D, Farber M, Ewend M, Marston W, Keagy B.
Complications of lumbar drainage after thoracoabdominal aortic aneu-
rysm repair. J Vasc Surg 2001;34:623-7.
8. Dardik A, Perler B, Roseborough G, Williams M. Subdural hematoma
after thoracoabdominal aortic aneurysm repair: An underreported com-
plication of spinal fluid drainage? J Vasc Surg 2002;36:47-50.
9. Acher CW, Wynn MM, Hoch JR, Popic P, Archibald J, Turnipseed W.
Combined use of cerebral spinal fluid drainage and naloxone reduces
the risk of paraplegia in thoracoabdominal aneurysm repair. J Vasc Surg
10. Huang TY, Chung HW, Chen MY, Giiang LH, Chin SC, Lee CS, et al.
Supratentorial cerebrospinal fluid production rate in healthy adults:
Quantification with two-dimensional cine phase-contrast MR imaging
with high temporal and spatial resolution. Radiology 2004;233:603-8.
11. Francia A, Parisi P, Vitale AM, Esposito V. Life-threatening intracranial
hypotension after diagnostic lumbar puncture. Neurol Sci 2001;22:
12. Gaucher DJ Jr, Perez JA Jr. Subdural hematoma following lumbar
puncture. Arch Intern Med 2002;162:1904-5.
13. Sciubba DM, Kretzer RM, Wang P. Acute subdural hematoma follow-
ing a lumbar CSF leak caused by spine surgery. Spine 2005;30:E730-2.
cerebrospinal fluid hypotension. Rev Esp Anestesiol Reanim 2002;49:
15. Dillon WP, Fishman RA. Some lessons learned regarding the diagnosis
and treatment of spontaneous intracranial hypotension. Am J Neurora-
16. Roll JD, Larson TC, Soriano MM. Cerebral angiographic findings of
spontaneous intracranial hypotension. Am J Neuroradiol 2003;24:
17. de Noronha RJ, Sharrack B, Hadjivassiliou M, Romanowski C. Sub-
dural Hematoma: a potentially serious consequence of spontaneous
JOURNAL OF VASCULAR SURGERY
Volume 49, Number 1
Wynn et al 33
intracranial hypotension. J Neurol Neurosurg Psychiatry 2003;74:
18. Kelsaka E, Sarihasan B, Baris S, Tur A. Subdural hematoma as a late
complication of spinal anesthesia. J Neurosurg Anesthesiol 2003;15:
19. Marini CP, Levinson J, Caliendo F, Nathan IM, Patel D, Faust G, et al.
Control of proximal hypertension during aortic cross-clamping: its
effect on cerebrospinal fluid dynamics and spinal cord perfusion pres-
sure. Semin Thorac Cardiovasc Surg 1993;10:51-6.
20. Kazama S, Masaki Y, Maruyama S, Ishihara A. Effect of altering cere-
brospinal fluid pressure on spinal cord blood flow. Ann Thorac Surg
21. Saether OD, Juul R, Aadahl P, Stroholm T, Myhre H. Cerebral hae-
modynamics during thoracic- and thoracoabdominal aortic aneurysm
repair. Eur J Endovasc Surg 1996;12:81-5.
22. Miyamoto K, Ueno A, Wada T, Kimoto S. A new and simple method of
preventing spinal cord damage following temporary occlusion of the
thoracic aorta by draining cerebral spinal fluid. J Cardiovasc Surg
23. Blaisdell F, Cooley D. The mechanism of paraplegia after temporary
thoracic aortic occlusion and its relationship to spinal fluid pressure.
24. Coselli JS, LeMaire SA, Köksoy C, Schmittling ZC, Curling PE. Cere-
brospinal fluid drainage reduces paraplegia after thoracoabdominal
aortic aneurysm repair: results of a randomized trial. J Vasc Surg
25. Svensson LG, Hess KR, D’Agostino RS, Entrup M, Hreib K, Kimmel
W, et al. Reduction of neurologic injury after high-risk thoracoabdomi-
nal aortic operation. Ann Thorac Surg 1998;66:132-8.
26. Weaver KD, Wiseman DB, Farber M, Ewend MG, Marston W, et al.
Complications of lumbar drainage after thoracoabdominal aortic aneu-
rysm repair. J Vasc Surg 2001;34:623–7.
ing lumbar spinal drainage during repair of thoraco-abdominal aneu-
rysm. Anaesthesia 2001;56:168-70.
28. Dardik A, Perler BA, Roseborough GS, Williams GM. Subdural hema-
toma after thoracoabdominal aortic aneurysm repair: An underreported
complication of spinal fluid drainage? J Vasc Surg 2002;36:47-50.
lar hematoma following thoracoabdominal aortic repair: an unreported
complication of cerebrospinal fluid drainage. Eur J Cardiothorac Surg
30. Foelholm R, Waltimo O. Epidemiology of chronic subdural haema-
toma. Acta Neurochir (Wein) 1975;32:247-50.
31. Kudo H, Kuwamura K, Izawa I, Sawa H, Tamaki N. Chronic subdural
hematoma in elderly people: present status on Awaji Island and epide-
miological prospect. Neurol Med Chir (Tokyo) 1992;32:207-9.
Submitted May 28, 2008; accepted Jul 25, 2008.
G. Patrick Clagett (Dallas, Tex). What was the temporal
relationship between operation and intracranial hemorrhage? Did
patients develop hemorrhage during operation or later in the
postoperative period? Was the spinal drain in place at the time of
hemorrhage and were there technical problems such as draining
too much CSF [cerebrospinal fluid] too fast?
Dr Matthew Mell. Most of the occurrences of bloody spinal
fluid drainage occurred during the procedure or shortly thereafter.
All patients had a functioning drain in place. No patients had
excessive fluid drained or exceedingly low pressures, as the drain-
age was performed manually, and therefore did not rely on a
passive drainage system.
Dr Starros Kakkos (Detroit, Mich). I have a couple of
questions. What is the mechanism of developing intracerebral
hematoma? Also, how much heparin do you give during the
Dr Mell. The mechanism of bleeding is not completely
understood, but it is thought to be due to the effect of pressure
changes on the dural sinuses or the cortical veins, which cross
the subdural space. These veins become engorged as the CSF
pressure decreases, increasing the risk for disruption and bleeding.
There may also be an effect of the rate of pressure change, with
sudden drops in the CSF pressure further increasing that risk.
With regard to heparinization, we perform open TAAA
[thoracoabdominal aortic aneurysm] repair without CPB [car-
diopulmonary bypass], and patients in this group were not
routinely heparinized. The patients undergoing endovascular
repair did receive a standard dose of heparin; for these patients
we have adjusted our protocol, aiming for a higher target CSF
Dr William Jordan (Birmingham, Ala). We have had some
challenges at our institution with collaboration from other
services. Specifically, the anesthesia team has hesitation that
placing the drain may create a spinal hematoma, or epidural
hematoma, leading to the neurologic problem we are trying to
avoid. You have demonstrated a wonderful series where you had
none, is that correct, you had no spinal drain complications in
Additionally, our neurosurgeons have great concern about the
intracranial pathology that might create complications, as you
demonstrated. You have elucidated some of the factors that you
would consider high risk. Would you then consider scanning
everyone with a CT prior to placing a drain, to pick up any of these
Dr Mell. To answer your first question, there were no spinal
hematomas in our series. We are very fortunate to work in very
close collaboration with anesthesiologists dedicated to optimal
operative anesthetic management during these complex cases,
which includes expertise in placement and monitoring of the CSF
drain. This dedication has been instrumental in the success of our
Optimal preoperative imaging for identifying intracranial
pathology is a difficult question to answer. The incidence of
asymptomatic subdural hematoma in the population increases
with age, although it remains quite uncommon even in elderly
patients. While it is true that some patients will have unrecog-
nized intracranial pathology at the time of surgery, we have not
adopted an approach of routine preoperative imaging. Rather,
we have selectively imaged patients considered to be at higher
risk for intracranial pathology. These may include patients with
mild dementia, a history of traumatic brain injury, alcohol
abuse, and so forth.
Dr Richard Cambria (Boston, Mass). I have always been
impressed at the great variability in volume of CSF drainage when
you set that bag at a passive 10 cm of water. So isn’t controlling the
volume just a question of what pressure threshold you will accept?
I think it is particularly important postoperatively.
Dr Mell. Thank you for your comments. We do not use a
technique of passive drainage with a preset pop-off valve, but
instead drain manually in 5-ml increments to achieve our target
CSF pressure. This technique controls both the volume and rate
of CSF drainage, factors we think are important to prevent
bleeding. With this approach, our average drainage is much
lower than that of other studies looking at CSF volume and
bleeding complications. Of course, we rely on the anesthesiol-
ogists to pay strict attention to operative drain management.
JOURNAL OF VASCULAR SURGERY
34 Wynn et al
Additionally, the newer drain is able to more precisely regulate
the CSF pressure, enabling the anesthesiologists to time the
drainage so that the target CSF pressure is reached just when we
are ready to cross-clamp. This ability to regulate has allowed us
to minimize the amount of drainage before the cross-clamp to
about 80 ml. Postoperatively, the ICU [intensive care unit]
nurses have been trained to continue the manual drainage until
the patient is awake with normal leg strength. These are some of
the techniques that we have been able to use control the amount
and rate of CSF drainage for these procedures.
COLLECTIONS OF PAPERS
On the Web version of the Journal, selected articles have been grouped together for the convenience of the
readers. The current collections include the following:
American Board of Vascular Surgery
Basic Science Reviews
Lifeline Research Meeting Abstracts
JOURNAL OF VASCULAR SURGERY
Volume 49, Number 1
Wynn et al 35