Atlas-of-Neurosurgical-Techniques-Spine-and-Peripheral-Nerves

95
InterbodyFusionandTranspedicularScrewFixation
689
Cloward's series! documented
"long-term"
follow-up
in
97 patients with
lytic
spondylolisthesis.
An excellent
or
goodoutcome was achieved in 95 patients,
and 94 patients
hadapparent fusions on radiographs.
Although
still
some-
what controversial,
there
is evidence
that decompression
and fusion are superior to decompression
alone in the man-
agement of symptomatic
spondyI0Iisthesis.!9
Although
it is
difficult to document
an obvious
superiority
of PLIF over
other fusion techniques for this condition,
PLIF with
pedicle
screwand plate constructs
may be ideal. The foraminal
and
central canal areas are restored, the anatomical
alignment
is
corrected, and the posterior
tension
band is re-created
by
the fixation construct.
Un et af14reported a 74% satisfactory
clinical
result and a
93% fusion rate in 71 patients
who
underwent
PLIF for
spinal stenosis. Similarly,
Hutter4
reviewed
142 patients
with spinal stenosis treated with
PLIF and obtained
good or
excellent results in 78% and a 91%fusion rate.
Unetaf14also reported the results of PLIFfor 143 patients
with recurrent disk herniations.
A satisfactory
outcome
was
found in 74%of patients and a solid fusion in 92% of patients.
Un's results for various indications
are listed in Table 95-2.
Branchand Branch2 reviewed
46 cases of PLIF for recurrent
disk herniations with a mean follow-up
of 12 months
and an
averageof 1.4 previous
operations.
A satisfactory
outcome
Table95-2
Posterior Lumbar Interbody
Fusion
Outcomes for Various Spinal ConditionsO
References
1. ClowardRB.Thetreatmentof ruptured intervertebral disks by verte-
bralbodyfusion,I: Indications, operative technique, after care.J Neu-
rosurg1953:10:154-168
2. Branch CL,Branch CLJr. Posterior
lumbar
interbody
fusion:
the key-
stone technique.
In: Lin PM, Gill K, eds. Lumbar
Interbody
Fusion.
Rockville,MD: Aspen: 1989:211-219
3. BrodkeDS,DickJC, Kunz DN. et al. Posterior
lumbar
interbody
fusion:
a biomechanical
comparison,
including
a new threaded
cage. Spine
1997:22:26-31
4. Hutter CG.Spinal stenosis and posterior
lumbar
interbody
fusion. Clin
Orthop 1985:193:103-114
5. LinPM. Posterior lumbar
interbody
fusion technique:
complications
and pitfalls. Clin Orthop 1985:193:90-102
6. Onesti ST,Ashkenazi
E. The Ray threaded
fusion cage for posterior
lumbar interbody fusion. Neurosurgery
1998;42:200-205
7. Steffee AD, Sitkowski
Dj. Posterior
lumbar
interbody
fusion
and
plates. Clin Orthop 1988;227:99-102
8. Kirkaldy-Willis WH, Farfan HE Instability
of the lumbar
spine. Clin
Orthop 1982:165:110-123
9. Frymoyer JW, Newberg
A, Pope MH, et al. Spine radiographs
in pa-
tients with low-back
pain. J Bone Joint Surg Am 1984:66:1048-1055
was documented
in 76% of patients,
which
compared
favor-
ably
with
the
outcome
in similar
patients
treated
with
laminectomy
alone
(40% satisfactory
outcome)
or
with
laminectomy
and lateral fusion (66% satisfactory
outcome).
If
PLIF can be done safely and quickly
with
satisfactory
fusion
rates, it can eliminate
the potential
for recurrent
operations
at that level and effectively
decompress the neural elements.
One possible disadvantage
of segmental
fusion in patients
with
degenerative
spinal conditions
is the potential
to accel-
erate degeneration
at levels adjacent to the fusion. There are
additional
mechanical
stresses placed on the adjacent motion
segments after spinal fusion. This can require further
lumbar
spine operations
to treat the adjacent-level
degeneration.
.
Conclusions
There is still controversy regarding the
use
of
fusion
in
the
management
of degenerative
conditions
of the lumbar
spine. However,accumulated clinical experience over more
than 4 decades seems to validate the PLIFtechnique as a
safe and efficacious procedure for a variety of conditions.
There is also evidence
that for recurrent
lumbar disk herni-
ations
or spondylolisthesis
it may be superior
to repeat
diskectomy
or decompression
alone. Generally, one can ex-
pect a 75% clinical
success
rate and fusion
rates>
90%.
Advantages of the PLIFtechnique include a familiar poste-
rior approach, decompression of the neural elements, and
restoration of disk space height with a load-bearing graft.
The addition of segmental fixation can provide immediate
postoperative
stability, correct anatomical
deformities,
and
enhance
fusion rates. Continued
improvements
in instru-
mentation
systems
provide
safer, more
rapid,
and more
standardized PLIFtechniques. Potential long-term
conse-
quences of segmental fusion such as adjacent-level degen-
eration must also be further investigated and characterized.
Prospective
studies
using
careful
patient
selection,
stan-
dardized
outcome
measures,
and long-term
follow-up
re-
view are necessary to identify patients
for whom PLIF
would be the optimal treatment
option.
10.
Sidhu KS, Herkowitz HN. Spinal instrumentation
in the management
of
degenerative
disorders of the lumbar spine. Clin Orthop 1997:335:39-53
11.
Branch
CLJr. The case for posterior
lumbar
interbody
fusion.
Clin
Neurosurg
1996;43:252-267
12.
Branch CLJr. Posterior
lumbar
interbody
fusion. In: Hardy RW Jr, ed.
Lumbar Disk Disease. 2nd ed. New York: Raven; 1993: 187-200
13.
Cautilli RA.Theoretical
superiority
of posterior
lumbar
interbody
fu-
sion. In: Lin PM, ed. Posterior
Lumbar Interbody
Fusion. Springfield,
IL: Charles CThomas:
1982:82-93
14.
Lin PM. Cautilli RA,Joyce ME Posterior
lumbar
interbody
fusion. Clin
Orthop
1983: 180:154-168
15.
Evans JH. Biomechanics
of lumbar fusion. Clin Orthop
1985: 193:38-46
16.
Stonecipher
T, Wright
S. Posterior
lumbar
interbody
fusion
with
facet-screw
fixation. Spine 1989;14:468-471
17.
Matsuzaki H,Tokuhashi Y,Matsumoto
F,et al. Problems and solutions of
pedicle screw plate fixation of lumbar spine. Spine 1990; 15: 1159-1165
18.
Lin PM. Technique
and complications
of posterior
lumbar
interbody
fusion.
In: Lin PM, Gill K, eds. Lumbar
Interbody
Fusion. Rockville,
MD: Aspen: 1989:171-199
19.
Turner JA. Herron
L, Deyo RA. Meta-analysis
of the results
of lumbar
spinal fusion. Acta Orthop Scand Supp11993;251
:120-122
Condition
No. of Patients
%Satisfaction
% Fusion
Recurrentdisk
149
74
92
Spinalstenosis
69
74
93
Instability
43
83
77
Narrowdisk
107
81
79
Lateraldisk
61
94
98
Midlinedisk
143
88
93
GAsreportedby Un et al.14

96
Posterior Fusionwith Hook-Rod
Instrumentation
Eldan Eichbaum, Cary Alberstone, and Bruce McCormack
.
Indications
.
Contraindications
.
Advantages
.
Disadvantages
.
Objective
.
Instrumentation Construct Design
.
Surgical Technique
Positioning
Incision and Dissection
Indications for spinal instrumentation
include traumatic,
degenerative, neoplastic, iatrogenic, and infectious etiolo-
gies of spinal instability. Spinal stability has been defined as
the ability of the spine to resist physiological loads and pre-
vent excessivemotion, spinal deformity, neurological injury,
and chronic pain.! Several theories of spinal instability have
been described.2-5 In clinical practice, the three-column
spinal theory of Denis,6based on a thoracic fracture model,
is the most widely accepted (Fig.96- t). Spinal instability is
defined as greater than two-column involvement.G.?
Spinal instrumentation affords immediate stabilization,
facilitating fusion and allowing for early patient mobiliza-
tion. It maybe placed ventrally, dorsally, or in combination,
depending on the mechanism of injury and the degree of
spinal column instability. Posterior instrumentation
is re-
quired when translational deformity is present, due to its
superior resistance (when compared with anterior instru-
mentation)
to torsional and translational
forces. Even
though this stabilization is immediate, the ultimate goal of
any instrumentation
construct remains the establishment
of a solid arthrodesis because failure to achieve a solid
arthrodesis will invariably lead to instrumentation
failure
(Fig.96-2).
Hook-rod instrumentation
is the most common poste-
rior construct
placed in the thoracic spine due to its
strength and versatility. Hooks may be placed under the
lamina, abutting the pedicle or encircling the transverse
process. Hook-rod constructs
employ multiple fixation
690
Hook Insertion
Laminar Hook Insertion
Pedicle Hook Insertion
Transverse Process Hooks
Coupling
Bone Graft Harvest
Fusion Preparation
Closure
.
Complications
.
Conclusions
points facilitating deformity correction, load sharing,and
pullout resistance.
.
Indications
Spinal instability or spinal deformity due to:
. Trauma
.
Neoplasm
. Infection
.
Osteoporosis
.
Paralytic scoliotic deformities
.
Ankylosingspondylitis fractures
.
Contraindications
.
Multilevel laminar, pedicle, and transverse process
fractures
.
Multilevel laminectomies
.
Lifeexpectancy of < 6 months
.
Incompetent anterior column
.
Spinal stabilization requiring decompression
. Severe spinal stenosis

96
Posterior Fusionwith Hook-Rod Instrumentation
691
~
,..-
Rgure 96-1
The three-column spinal theory of Denis divides the spinal
columninto three segments: anterior, middle, and posterior columns.
c
:..0
=.
'"=
gt",
-~
Cc-.iii_
'"c
=.--
ua,
: e
--
CIJ.
o
Bone Fualon
Component
Spina.
Implant
Component
r
Time
Surgery
Figure96-2
The relationship between bone fusionacquisition and
spinalimplantintegritychangesas afunctionoftime followingsurgery.
. Advantages
. Lowprofile
. Biomechanicallysuperior
. Resistsbending, shearing, and rotational forces
. Correctsspine deformities at various planes
. Nosignificant risk of catastrophic vascular injury
. Inexpensive
. Excellentfor severe osteoporosis
. Minimaldevice pullout due to multiple fixation points
. Morbidlyobese or pregnant patients may be placed in
lateraldecubitus position
. Lowerrisk of neurological injury (pedicle and transverse
process hook)
. Multilevelfixation allows for load sharing
. Maybe used in multilevel laminectomy
.
Disadvantages
.
Requiresmultilevel fixation
.
Riskof dural laceration
.
Cerebral spinal fluid fistula
.
Neural element injury
.
Laminar preparation time-consuming and tedious
. Riskof neural element injury (sublaminar hooks)
. Poor control in axial plane
.
Instrumentation
failure due to hook pullout or pedicle
hook fracture
. Objective
Use biomechanical forces to correct spinal deformity, main-
tain spinal alignment, and create a stable environment to
promote solid arthrodesis
.
Instrumentation ConstructDesign
The nuances of construct design are beyond the scope of
this chapter, but some principles should be discussed. Pre-
operative construct design includes determination of hook
placement and orientation, the spinal levels to be fused,
and the levels to be fixated.
The determination of hook placement is based on several
variables, such as the spine region, pathology, biomechanical
forcesdesired, and surgeon preference. Hooks may be placed
under the lamina, abutting the pedicle, or encircling the
transverse process. In general, sublaminar hooks cephalad to
the pathological level (or deformity) face rostrally, whereas
sublaminar hooks distal to the pathological level face cau-
dally. Pedicle hooks face rostrally and transverse process
hooks face caudally due to the orientation of the thoracic
facet joints and transverse processes, respectively. In addi-
tion, transverse process hooks are frequently associated with
an adjacent pedicle hook, forming a claw configuration.
Aclaw configuration is the strongest type of hook construct,
whereby two ipsilateral hooks faceone another at either the
same or consecutive level(s). These "claws" are most com-
monly placed at the terminal ends of a construct (Fig.96-3).
Determination ofthe spinal levelsto be fixated and fused is
also planned preoperatively. Instrumentation fixation most
commonly extends three levels above and two levels below
the pathological level, creating equal cephalad and caudad
fixation.Thisisdue to the caudal position ofthe posterior ele-
ments relative to the vertebral bodies. The multilevel fixation
facilitates deformity correction and decreases instrumenta-
tion failure.The fusion mass may cover the pathological seg-
ments only or span the entire length of the construct. It isthe
authors' opinion that a fusion/instrumentation mismatch is
optimal, whereby the fusion mass covers only the pathologi-
cal segments. This technique allows for removal of the long
instrumentation construct after a solid arthrodesis has oc-
curred, minimizing the risk of complications, such as facet
arthropathy, to normal motion segments.

692
LumbarandLumbosacralSpine:PosteriorApproach
A
B
C
Figure 96-3
Asingle-level claw configuration is most commonly placed at the terminal ends of a construct.
.
SurgicalTechnique
Positioning
After induction of anesthesia, the patient is carefully placed
in the prone position. This should be done with adequate
assistance, especially when gross spinal instability is pre-
sent. Appropriate padding is provided for the hips, the
brachial plexus, and the extremities.
The abdomen
is
allowed to fall ventrally to decrease intra-abdominal pres-
sure and epidural venous engorgement and thus to mini-
mize epidural blood loss.
Incision and Dissection
Agenerous dorsal incision is drawn allowing for adequate
rostral, caudal, and lateral exposure. The incision may be
midline, paramedian, or midline with lateral extension (Le.,
hockey stick) depending on the approach desired and the
extent of ventral exposure required (Fig.96-4). Anincision
is also drawn over the iliac crest unilaterally or bilaterally
for autologous bone graft harvest. The Bovieand periosteal
elevator are used to dissect the paraspinous musclesfrom
medial to lateral, exposing the spinous processes, laminae,
facet joints, and transverse processes. Care is taken not to
violate the facet capsule with the Bovie,especiallyin nonin-
strumented segments. This can be accomplished by either
careful Boviedissection over the capsule or blunt dissection
using a pair of periosteal elevators and sponge.
Frequent intermittent release of the self-retaining retrac-
tors is performed to decrease muscle ischemia and mini-
mize muscle necrosis, postoperative infections,and wound-
healing difficulties. Cell-saver suction is recommended,
except in patients with infection or cancer. Hemostasismay
be obtained with bipolar cautery, Gelfoam,or SurgiceLBone
wax is used sparingly due to its inhibition of fusion.

...
96
Posterior Fusionwith Hook-Rod Instrumentation
693
A
Figure96-4
Different skin incisions for dorsolateral exposure of the
thoracic and lumbar spine. (A) Dorsal midline incision. (8) Semilunar
incision.(C) Hockey-stick incision. (D) Straight paramedian
incision.
Anintraoperative plain radiograph is obtained to assure
thecorrect spinal level. The authors prefer an anteroposte-
rior (AP) radiograph for localization in the upper and
midthoracic spinal column, and a lateral radiograph for lo-
calizationin the lower thoracic region. Ribsare counted in
the AP projection, whereas
vertebral
bodies from the
sacrumare counted on the lateral projection. C-arm fluo-
roscopyis used in cases involving pedicle screw constructs
butmayalso be used for intraoperative localization. Spinal
decompression. if required. should be performed prior to
dorsalelement preparation.
HookInsertion
Inthe authors' opinion, the optimal position for hook inser-
tionis under the lamina. Sublaminar hooks have the great-
estbone-metal interface and the strongest purchase. This
decreases the risk of hook pullout or migration and allows
forexcellent deformity correction and stabilization of the
spinal column. lumbar or sacral sublaminar hook place-
ment is well accepted. whereas thoracic sublaminar hook
insertionrostral to the conus medullaris is more controver-
sialbecause of the relatively narrow thoracic spinal canal
andits tenuous blood supply.
laminar Hook Insertion
Laminarhook insertion is initiated with laminotomies of
the rostral half of the lamina below and the caudal half of
B
Figure 96-5
Arectangle-shaped semilaminectomy iscreated byre-
movingthe ligamentum flavumand medialfacet withangled curettes
and Kerrisonrongeurs.
D
the lamina above the level of hook placement. In sublami-
nar hook insertion. removal of the spinous process to its
base is frequently necessary to allow for hook and rod cou-
pling. The ligamentum fIavum and medial facet are re-
moved with angled curettes and Kerrisonrongeurs. creating
a rectangular laminotomy (Fig. 96-5). This laminotomy
must be large enough to allow for adjacent contralateral
hook placement. but excessive bony removal may weaken
the lamina for hook placement or preclude a claw configu-
ration on the same lamina.
The widest possible hook should be inserted. creating the
greatest bone-implant interface and thereby decreasing the
risk of hook pullout. Multiple hook trials, or a laminar
tester, may be used to choose the correct hook by measur-
ing the lamina size and contour. Ifthe hook does not fit the
contour of the lamina, then either the lamina may be fur-
ther contoured or another hook may be inserted. Coupling
the hooks to the rod may be performed after placing all
hooks initially, or in a sequential tightening technique,S as
will be described here.
Initially, the rod is measured, cut. -and contoured to fit
the spinal curve. The first hook is secured at either termi-
nus, and then each successive hook is sequentially coupled
to the rod until the unilateral construct is complete. Dor-
sally directed forces are applied at all times to avoid ventral
migration of the hooks. Each hook is securely tightened
upon insertion and compressed onto its respective lamina
(Fig. 96-6).
Distractive or compressive
forces may be
placed sequentially or after completion of coupling. This
step is repeated for the contralateral rod. The construct is
completed by the placement of two cross-fixators. Cross
fixation increases the stability of the implant by resisting
torsional and translational
forces as well as minimizing
hook pullout.

694
Lumbarand LumbosacralSpine: PosteriorApproach
A
~
\
\Y"--"'--'"
'\
B
Figure 96-6
The measured and cut rod iscontoured to fitthe spinalcurve.(A)Thefirsthook issecured at the terminusand eachsuccessivehook
issecuredsequentiallyuntilthe unilateralconstruct iscomplete.(B)Eachhookistightened and compressed into its respectivelamina.
Pedicle Hook Insertion
The most common site for hook insertion in the thoracic
spine is the pedicle, resting caudad to the inferior facet and
cephalad to the rostral facet. The coronal orientation of the
thoracic facetjoints promotes this placement, whereas the
sagittalorientation of the thoracolumbar facetjoints preclude
it. Furthermore, this lateral placement avoids spinal cord
compression,minimizing the chanceof neuralelementinjury.
The inferomedial
quadrant
of the caudal facet is re-
sected by either a high-speed drill or an osteotome(Fig.
96-7).
The hook is inserted pointing
rostrally, engaging
the pedicle. Maximal
pedicle purchase is obtained with
adequate bone resection.
Care is taken not to perform
Figure 96-7
Pedicle hook insertion. Dorsal view of the thoracic facet demonstrating
the inferior articular process. (A) Before and (B)after bone
removal
for placement
of a pedicle hook.

r
96
Posterior Fusionwith Hook-Rod Instrumentation
695
Figure96-8
(A) Lateralview of pedicle hook in place demonstrating appropriate bone removal. Note how the bifid hook properly engagesthe
pedicle.(8) If too much bone isremoved, the pedicle may migrate rostrally and cut into the pedicle or toggle out of position. (C) If too little bone is
removed,the hookwill not engagethe pedicleand mediolateral hook migration may occurwith manipulation of longitudinal members.
excessiveor insufficient bone removal because this may
leadto pedicle fracture or inadequate pedicle purchase,
respectively(Fig. 96-8).
Bifid and nonbifid hooks are
available,each possessing excellent purchase due to the
closeapproximation of the inferior and superior
facet
complex. Bifid hooks are designed
to engage
the pedicle,
therebyimproving its purchase, but aggressive insertion
maycausepedicular fracture and subsequent weakening
oftheconstruct.
Transverse Process Hooks
Transverseprocess hooks are also commonly
placed in the
thoracic spine
due
to
the
relatively
large
transverse
processes.
The ventrorostral
orientation
of these transverse
processespromotes hook insertion, most commonly in the
caudaldirection.
These hooks are inserted
at the cephalad
endof the transverse
process after stripping
off the costo-
transverseligament
with
a sharp dissector. A claw configu-
ration is created when
combined
with
a pedicle hook. The
transverseprocesses in the thoracolumbar
region are rudi-
mentary precluding
transverse process hook insertion.
Coupling
Securing the hooks to the rod is termed coupling. The hooks
maybe coupled to the rod after all of the hooks are in-
serted,orin a sequential manner. Hooks that accept the rod
ontop are "top-loading,"
whereas hooks that engage the
rod on the side are "side-loading."
When utilizing
top-
loadingsystems,all the hooks are inserted initially followed
byplacementof the rod. When utilizing side-loading sys-
tems,all hooks may be inserted initially and then coupled
tothe rod, or sequentially, as previously described.
BoneGraft Harvest
Thebone harvested during
preparation
of the spine may be
used as graft material, but iliac crest cortical and cancellous
boneis of superior quality. The caudal length of the incision
dictateswhether
a separate
incision is required for the graft
harvestsite. Once a large portion
of the dorsal
iliac crest is
exposed and
all
soft
tissue
removed,
thin
unicortical
(matchstick) strips are created with
a straight
osteotome
(or
oscillating
saw)
and
mallet.
These
sticks
are
carefully
removed, and the exposed cancellous bone is gouged and
collectedinto a container with normalsalineor collected
blood. Care is taken not to cut too deeply into the ilium and
violatethe deep corticallayer.Followingcompletionof the
harvest,copious amounts of antibiotic fluidare irrigated
into the wound and a large piece of Gelfoam is placed
firmly on the cancellous bone. After hemostasis is obtained,
the graft site wound is closed in layers.
Fusion Preparation
The fusion bed is prepared using a high-speed air drill to
decorticate
to
healthy
bleeding
bone.
The
transverse
processes
are most commonly
decorticated
in the thoracic
spine due to their large size, strength, and distance from the
spinal canal. The lamina and facet complexes
are also excel-
lent recipients of bone graft but frequently are not available
due to bony destruction from either or both the pathologi-
cal process and the decompression. Decortication
should
not be too extensive so as to weaken the recipient bone.The
wound is copiously irrigated
with 2 to 3 L of antibiotic
solution using a pulsed irrigation suction device, prior to or
following decortication.
The
harvested
hip
cancellous
bone
and
spinous
process-lamina
nonstructural
cortical bone graft are morcel-
lized and kept separately. Many surgeons combine this graft
material with substances containing bone morphogenic pro-
tein to increase osteoblast
production
and enhance
fusion
rates. These substances
also improve the handling ability of
the graft material. This graft material
is placed bilaterally,
spanning the desired levels offusion. The harvested iliac bone
is placed initially, followed by the spinous process-lamina
bone.Ideally,the unicortical strips areplaced with their can-
cellous half ventral, spanning at least one motion segment.
Closure
Woundclosureis initiatedwith releaseofthe retractorsand
meticulous paraspinous muscle hemostasis. Prior to closure,
large drains may be placed in the subfascial space tunneled
through a distant stab incision, especially if hemostasis is
not optimal. Largeabsorbable suture is used for fascial clo-
sure and subsequent layers up to the skin (Le.,Vicryl0, 00).
The skin may be secured with either staples or a monofila-
ment suture.

696
Lumbarand LumbosacralSpine: PosteriorApproach
.
Complications
Complicationsassociated with dorsal hook-rod instrumen-
tation systems include hook pullout, neural element injury
due to hook insertion, or construct failure.
Hookpullout most frequently occurs at the caudal termi-
nus of a construct, especially in constructs using three-
point bending forces to reduce deformity. Aclaw configura-
tion increases the bone/implant
interface and construct
strength, decreasing the risk of hook pullout and the motion
segments immobilized. Therefore, this configuration should
be employed at the termini of these constructs to minimize
this complication.
Sublaminar hooks are at risk for neural element compres-
sion on either insertion or in constructs under certain bio-
mechanical forces (Le., three-point bending). Sublaminar
hook insertion must be performed under a meticulously
prepared lamina, always adhering to the ventral laminar
surface. Furthermore, preoperative imaging should alert the
surgeon to the presence of spinal stenosis, precluding the
placement of sublaminar hooks.
Pedicle hooks may also migrate and encroach on the
neural elements in the spinal canal or nerve root in the
References
1.
White M. Punjabi MM.Clinical Biomechanics of the Spine. Philadel-
phia:
JB Lippincott;
1978
2.
Benzel Ee. The anatomic basis of spinal instability. Clin Neurosurg
1994;41 :224-241
3.
Farfan
HF. Gracovetsky
S. The nature
of instability.
Spine
1984;9:
714-719
4.
Posner
I. White
M
III. Edwards
WT. Hayes We. A biomechanical
analysis
ofthe
clinical stability
ofthe
lumbar
and lumbosacral
spine.
Spine 1982;7:374-389
neural foramen. Meticulous bony removal to assure tight
fitting of the hook between the facet complexesdecreases
this risk. Bifidpedicle hooks may fracture the pedicle.ren-
dering it (and the construct) weaker. Multiplehooktrialsin
the facet complex should decrease the incidence of this
complication.
.
Conclusions
In summary, dorsal hook-rod instrumentation isindicatedin
individuals with spinal instability, especially with transla-
tional deformity. Dorsal hook-rod instrumentation con-
structs afford superior strength when comparedwithventral
constructs. The three regions forhook insertionareunderthe
lamina, abutting the pedicle, or encircling the transverse
process. Due to the relatively narrow spinal canaland the
tenuous blood supply of the thoracic spine, manysurgeons
prefer not to utilize sublaminar hooks. However,in theau-
thors' opinion, sublaminar hooks may be placed safely
throughout the thoracic spine affordingthe greateststrength
for deformity correction, stability,and pulloutresistance.
5.
James KS. Wenger
KH. Schlegel JD. Dunn HK. Biomechanical evalua-
tion
of
the
stability
of
thoracolumbar
burst
fractures.
Spine
1994;19:1731-1740
6.
Denis F.The three column
spine and its significance
in the classifica-
tion of acute thoracolumbar
spinal injuries. Spine 1983;8:817-831
7.
Denis F. Spinal instability
as defined
by the three-column
spine con-
cept in acute spinal trauma.
Clin Orthop
1984;189:65-76
8.
Benzel EC.Ball PA, Baldwin NG. Marchand
EP.The sequential hook in-
sertion
technique
for universal
spine
instrumentation
application:
technical
note.J Neurosurg
1993;79:608-611

97
PosteriorFusionwith Luque Instrumentation
EldanEichbaum.CaryAlberstone.and BruceMcCormack
. Indications
. Contraindications
. Objectives
. Advantages
. Disadvantages
. SurgicalTechnique
Positioning
IncisionandDissection
Wires,Cables,Rods,andRectangles
The development of spinal instrumentation
systems
pro-
videsfor immediate spinal stability
and deformity
correc-
tionand promotes fusion and early mobilization
of patients.!
Theneed for postoperative bracing is minimized.
Prior to the
development of modern universal
spinal instrumentation
systems,2,3Luque popularized
a technique
using sublaminar
wire-rod fixation over multiple spinal levels. The ease of rod
contouring and multilevel
fixation
utilizing
Luque's tech-
niquepermits rigid spinal fixation.4 Deformity correction
is
obtainedby using the instrumentation
to apply three-point
bendingforces to the spine and thereby correct kyphosis.
Moreversatile or universal spine systems that utilize hooks
and screws were developed after Luque published
his tech-
niqueand are now the preferred spinal implant. Situations re-
main,however, in which the Luque wire-rod
construct
is the
ideal constuct. Paralytic scoliotic deformities
that
require
lengthyfixation from the upper thoracic spine to the ilium are
bestmanagedwith Luque wire-rod fixation.4 The authors pre-
ferthe Luquewire-rod fixation technique for the treatment
of
instabilityat the cervicothoracicjunction due to cancer. In
these patients osseous union is often not expected
and the
Luqueimplantmay loosen but is less likely to fail compared
witha screw-plate construct. The Luque system has a low pro-
fileandispreferredover some of the bulky universal spinal
implantsifthere is a lack of soft tissue coverage over the spine.
SublaminarWire PassagePreparation
SublaminarWire Passage
Fixationof Rodto lamina
Reductionof SpinalDeformity
SpecialConsiderationsandTechnique
FusionPreparation
Closure
.
Complications
.
Conclusions
.
Indications
Spinalinstabilityor spinalsagittaldeformitydueto:
. Trauma
.
Neoplasm
. Infection
.
Osteoporosis
.
Fixation of ankylosing spondylitis fractures (in this sub-
set of patients, osseous union is often not expected, and
the Luque implant may loosen but it is less likely to fail
than a screw-plate construct)
.
Contraindications
.
Multilevel laminar fractures
.
Multilevel laminectomies
.
Incompetent anterior column
.
Spinal stabilization requiring distractive or compressive
corrective
forces
697

698
Lumbarand Lumbosacral Spine: Posterior Approach
.
Objectives
. Use three-point bending movement forces to correct
spinal deformity and maintain spinal alignment
.
Promote solid arthrodesis
.
Advantages
As compared with anterior spinal instrumentation:
. Posteriorinstrumentation constructs are biomechani-
callysuperiortoanteriorshortsegmentconstructs.
. Resistingbending,shear,and rotationalforces
. Correctspinaldeformityinvariousplanes
. Nosignificantriskofcatastrophicvascularinjury
As compared
with
other
posterior
instrumentation
techniques
. Relativelyinexpensiveand simpleto apply
. Versatile(maybe applied throughout the entire length
ofthe spine)
. Excellentdeviceforuse insevereosteoporosis
. Multiplefixationpointsminimizedevicepullout.
. Maybe placed(with relativeease)in the lateraldecubi-
tus positioninmorbidlyobeseor pregnantpatients
.
Disadvantages
. Requiresmultilevelfixation,eliminating multiple nor-
malmotionsegments
. Riskofneuralelementinjury
. Riskofdural lacerationand cerebralspinal fluidfistula
. Laminarpreparationtime-consumingand tedious
. Poor control in the axial plane (requires competent
anterior column)
.
SurgicalTechnique
Positioning
Afterinduction ofanesthesia, the patient is carefully placed
in prone position.Thisshould be done with adequate assis-
tance, especiallywhen grossspinal instability is present. Ap-
propriate paddingisprovidedfor the hips,the brachial plexii,
and the extremities. Theabdomen is allowed to fallventrally
to decrease intra-abdominal pressure and epidural venous
engorgement, and thus to minimizeepidural blood loss.
Incision and Dissection
Agenerous dorsal incision is performed to provide adequate
rostral and caudal exposure. The incision may be midline,
paramedian, or midline with lateral extension (Le.,hockey
stick) depending upon whether a posterolateral approachis
needed to perform ventral cord decompression (Fig.97-1).
An incision is also made over the iliac crest for autologous
bone graft harvest. Monopolar cautery is used forhemosta-
sis and subsequent dissection. The Bovie and periosteal
elevator are used to dissect the paraspinous musclesfrom
medial to lateral, exposing the spinous processes,laminae,
facet joints, and transverse processes. Care is taken not to
violate the facet capsule with the electrocautery in nonin-
strumented segments. This is accomplished by either Bovie
dissection over the capsule or blunt dissection usinga pair
of periosteal elevators and sponge.
Frequent intermittent release of the self-retainingretrac-
tors decreases muscle ischemia, minimizing musclenecro-
sis, postoperative infections, and wound healingdifficulties.
Cell-saver suction is recommended except in patients with
infection or cancer. Boneand epidural bleeding maybebrisk
but most often occurs with a slow persistent ooze.Bonewax
is used sparingly if spinal fusion is being performed.
Gelfoamor Surgicelis preferred forbony hemostasis.
An intraoperative
plain radiograph is performed if no
landmark is present to assure the correct spinal level.The
authors prefer an anteroposterior (AP)radiograph to local-
ize the level in the upper and midthoracic spinal columnby
counting ribs in the AP projection. A lateral radiographis
used to localize the correct level in the lower thoracicspine
by counting vertebral
bodies from the sacrum. (-arm
flouroscopy is used for pedicle screw fixation inthe thoracic
spine, but may also be used for localization. Intraoperative
radiographs are correlated with surgical findings,preopera-
tive radiographs, and other imaging modalities such as
computed tomographic (CT)or magnetic resonanceimaging
(MRI)scans.
Wires,
Cables,
Rods, and Rectangles
Wires or multistrand stainless steel cables are passedunder
the lamina to fixate the rods to the spine. Wires (16,18,or
20 gauge) are available in single-beaded strands or openor
closed loops in varying lengths.s Surgeon preference and
availability dictate whether wire or cable is implanted.(a-
ble implantation requires additional instruments forsecur-
ing the cable including a top hat crimp, a tension/crimper,
and a torque wrench.
Stainless steel rods are either straight, prebent in an
L-configuration. or closed rectangles of various lengthsand
configurations. The rod diameter is either 4.8 mm (1/16in.)
or 6.4 mm (1/4in.) with varying lengths depending on its
configuration. Straight rods are 60 em (23.6 in.) in length;
prebent rods are 20, 40, and 60 em long with short armsof
3 or 5 em; and rectangles are 2 em wide and vary in length
from 4 to 50 cm.s
Sublaminar Wire Passage Preparation
The lamina are prepared for wire passage after adequate
spinal exposure. The interspinous ligament above andbe-
low each lamina to be instrumented
is removed with a
rongeur. The ligamentum flavum on the undersurface of
the lamina is removed by placing either a straight or an

97
Posterior Fusionwith Luque Instrumentation
699
A,a
C-E
Figure97-1
Different skin incisions for dorsolateral exposure of the thoracic and lumbar spine. (A) Dorsal midline incision. (8) Semilunar incision.
(e) Hockey-stickincision. (D) Straight paramedian incision. (E) .T-incision.
angledcurrette in the the epidural space and scraping the
tissue out from under the lamina. It is important that the
scraping is performed with a dorsally directed force to
avoid neural injury. Further bony removal is completed
with Kerrison rongeurs. The amount of bone removed
should be adequate to allow for safe wire passage but not
soextensive as to compromise the amount of lamina avail-
ablefor wire purchase. Sublaminar soft tissue is removed
toease wire passage and minimize risk of neural injury. A
Woodson(dental tool) elevator is used to palpate the sub-
laminarspace prior to passage of the wire(s). This serves as
a final inspection of the sublaminar space and its patency
forwire passage.
Sublaminar Wire Passage
The wire (or cable) may be placed under the lamina by
hand or by using a heavy needle holder. Asingle sublami-
nar pass is preferred, which is facilitated by bending the
wire in half, or in the case of a cable, using a double cable
with a wire connector (double leader). The wire tip is con-
figured with a wide radius of curvature, subtending 90 to
120degrees of arc (Fig. 97-2), and held at midshaft per-
mitting controlled passage but prohibiting forced entry.s
Thewire is placed under the caudal end of the lamina in
the midline, and carefully threaded in a cephalad direction,
with the tip following the dorsal surface of the lamina at all
times. If any impedence is encountered
during passage,
forced entry should be avoided and further dissection is re-
quired. The wire may be pulled
back and redirected
slightly, and if impedence to passage continues, the wire is
removed and the sublaminae space is reinspected with an
instrument (Le.,Woodson elevator, Penfield no. 4) and fur-
ther removal of ligament or bone is performed with a
Kerrison rongeur or curette. Once the sublaminar space is
cleared, the wire pass is reattempted. As the wire tip ap-
pears at the cephalad end of the lamina, its continued pas-
sage is facilitated by initially grasping it with an instru-
ment, such as a nerve hook or Woodson elevator, and
pulling upward until a sturdy needle holder may secure the
wire and complete the passage. The wire is then cut with a
wire cutter to create two equal halves. Each half is pulled
laterally
and temporarily
secured by either
pinching
(Fig. 97-3) or crossing the wire (Fig. 97-4) so as to con-
form to the lamina and minimize risk of dural compression.
Cables are temporarily secured to the lamina by pulling the
cut wire end through the closed loop end and using a
clamp to prevent ventral migration.
Wire cut edges are sharp and may easily cut through
gloves or skin and the surgeon and assistant(s) should be
cautious. Passage of bilateral double-stranded
wires may
be performed if the adequacy of strength of the construct
is a concern using only a single strand of wire to secure
the rod.

700
lumbarandlumbosacral Spine:PosteriorApproach
A
~
~
B
Figure 97-2
(A) Asingle sublaminarwire
pass is preferred. Tofacilitate this. the wire is bent in half and the tip is configured with a wide radiuswith
a 90 to 120 degree arc. (8) The wire is placed midline under the caudal end of the lamina and threaded cephalad with the tip followingthe dorsalsur-
face of the lamina at alltimes.
c
{Z~
- "}
Ie.
6
!J
.'
~
Q "
~
l
A-'_ i;..:./
)
Figure 97-3
Wire impingement
onto the dural sac is illustrated on
the left. This potential for neural compromise is present from the time
the wire is passed underneath the lamina until it is tightened
(twisted)
around the rod (demonstrated
on the right). The wire may be crimped
around the lamina to minimize its chance for causing injury during this
aspect of the operation (center with arrows).
Figure97-4
Thewiresare crossedand firmlyconformedtothedor-
sum of the lamina preventing any inadvertent migration into the
canal.

-
Fixation of Rod to Lamina
97
Posterior Fusion with Luque Instrumentation
701
Therodiscut to the appropriate length with a bolt cutter,
contouredto the desired conformation. and then placed on
thelamina.Therod is secured to the bone by twisting the
wirewitha heavywire twister until it does not move on the
rod,andthen the wires are cut to -3 to 4 em. After the ini-
tialtighteningis performed at each segment, loosening fre-
quentlyoccursat earlier twisted segments; therefore, the
wiresmayrequire sequential twisting to rigidly secure the
rodLOthe spine.The wires are cut to 1 to 2 cm after final
tighteningandbent mediallyor laterally (Fig.97-5).
Whenusing an L-rod, the shorter limb should be long
enoughsoas to rest on the contralateral facet complex, de-
creasingthe risk of rotation into the canal. In addition, the
shorterlimbof the rod should be placed through the
supraspinousligament and under the longer limb of the
contralateralrod.The wire at that segment should encom-
passbothrods.Occasionally,a portion of the caudad or
cephaladspinous process is removed to accommodate the
horizontallimbofthe L-rodor rectangle.
Whenusingcable,onemustpassthe cut leaderwireend
throughthe preformedloop on the opposite end of the wire
andpullsnuglyonto the rod and lamina. A"top hat" crimp is
threadedoverthe wire to the junction of the wire and pre-
formedloop.The leader wire is then threaded through the
Figure 97-5
After initial placement
of the wires, each segment
is
sequentially twisted
and tightened
to rigidly
secure
the
rod to the
spine.The wires are cut to 2 cm after the final tightening
and bent
mediallyor laterally.
tensionmechanismofthe tension/crimper.Thetophat crimp
is securely held in the jaws of the tension/crimper while a
torque wrench is applied to the tension/crimper to tighten
the wire to a uniform tension, usually 10 in.-Ib.4When the
desired tension is achieved, the top hat is crimped by tight-
ening the jaws of the tension/crimper, and the excess wire is
cut at the junction ofthe wire and top hat crimp (Fig.97-6).
Ifa single cable has been passed,a small connecting bar must
be passed over both ends ofthe wire to allow the formation
of a noose. Of note, crimping of the top hat should be
performed only after the segments have been sufficiently
tightened just prior to crimping. This decreases the risk of
loosening at a previously crimped segment. Ifloosening oc-
curs at a previously crimped site with the wire already cut,
then it is virtually impossible to be further tightened. In de-
formity correction, this may be avoided by tightening all the
cables and then crimping each cable successively,beginning
with the segment where the rod is furthest from the spine.
Reduction of Spinal Deformity
The Luque rod system may be used to reduce kyphotic de-
formity by leveraging the two Luquerods over the kyphotic
deformity. The shorter arms of two L-rods are placed
through
the most cephalad
and
caudad
interspinous
segments. Reduction of the deformity is accomplished by
Figure 97-6
When wires are sufficientlytightened, the top hat of
the wires are crimped by tightening the jaws of the tension/crimper
and the excesswireiscut.

702
Lumbar and Lumbosacral Spine: Posterior Approach
sequentially tightening the wires on both rods simultane-
ously beginning at the shorter arm end of each rod. This
maneuver will apply a three-point bending moment force
to straighten the spine (Fig. 97-7). The device cannot pro-
vide distractive or compressive forces to the spine due to
the wires sliding up and down the length of the rods and is
a major disadvantage of this system.
SpecialConsiderations and Technique
Wire fixation of the rods to the spine may be performed at
the base of the spinous process. This technique eliminates
the sublaminar passage of wire, thereby improving safety at
the expense of construct strength. This may be performed
only with a straight or L-rodbecause rectangles are not able
to approximate to the base of the spinous process. Drum-
mond and Keene described this techniqe using Wisconsin
"buttons" to facilitate wire-rod
fixation to the spinous
process.6Briefly,a hole is created in the base of the spinous
process with either a curved awl, a towel clamp, or even a
small round bur on a high-speed drill. Awire is bent in half
and the two free ends are threaded through two of the
three holes in the Wisconsin button. The wire ends are then
Figure 97-7
Thethree-point bendingcorrectiontechnique corrects
a kyphoticdeformityutilizingthe LuqueL-rodsystem.
Figure 97-8
Spinous process wiring utilizing Wisconsin buttons.
passed into the hole in the spinous process exitingthrough
the third hole of the contralateral button. Thisis repeated
on the contralateral side in the opposite direction,andthe
rods are secured to the base of the spinous processbytwist-
ing the wires as decribed previously (Fig.97-8). Thistech-
nique is especially useful in patients with ankylosing
spondylitis and sheer fractures precluding sublaminarwire
passage.
Fusion Preparation
The fusion bed is prepared using a high-speed air drillto
decorticate
to healthy
bleeding bone. The transverse
processes are most commonly decorticated in the thoracic
spine due to their large size, strength, and distancefromthe
spinal canal. The lamina and facet complexes are excellent
recipients of bone graft, but frequently are not availabledue
to bony destruction from either or both the pathological
process and the spinal canal decompression. Decortication
should not be so extensive as to weaken the recipientbone.
The wound is then copiously irrigated with 2 to 3 Lofan-
tibiotic solution, preferably using a pulsed irrigationsuction
device.
The
harvested
hip
cancellous
bone
and spinous
process-lamina
nonstructural cortical bone graft are then
morcellized and kept separately. Graft material may be
combined with substances containing bone morphogenic
protein (e.g., Grafton, Osteofil), to increase osteoblastpro-
duction and enhance fusion rates. These substancesalsoim-
prove the handling ability of the graft material. Thisgraft
material is then placed bilaterally,spanningthe desired
levels of fusion. The harvested hip bone is placed initially

97
PosteriorFusionwith LuqueInstrumentation
703
becausethis is the superior graft material, followed by the
spinousprocess-lamina bone. Ideally, the unicortical strips
are placed with
their cancellous side against the spine.
spanningat least one motion segment.
Closure
Wound closure
is then
initated
with
release of the retrac-
tors and meticulous
paraspinous
muscle hemostasis.
Some
surgeons prefer to place large drains in the subfascial
space
tunneled through
a distant
stab incision.
Large absorbable
suture is used for fascial closure and susequent
layers up to
the skin (Le.. Vicryl
O.00). The skin may be secured with
ei-
ther staples or a monofilament
suture.
.
Complications
Nerveinjury or dural laceration during wire passageis a
complication
specific to the Luque system that deserves spe-
cial mention. There is a 1 to 3% risk of nerve injury
with
sub-
laminar
wire
passage. Because the Luque system
requires
multiple
wire
passes into the spinal
canal, cumulative
risk
may be significant.
The surgeon should be aware of the spinal
canal dimensions
to determine
whether
it is safe to proceed
with sublaminar
wire
passage. Neurological
complications
References
1. Stillerman
CB,Gruen JP,Roy R.Thoracic and lumbar fusion: techniques
for posterior stabilization. In: Menezes AH, Sonntag VH, eds. Principles
ofSpinal Surgery. New York: McGraw-Hili;
1996:1199-1223
2.
Luque ER,Cardoso A. Segmental correction of scoliosis with rigid in-
ternal fixation:
a preliminary
report. Orthop Trans 1977; 1:136
3.
Luque ER,Cassis N, Ramirez-Wiella G. Segmental spinal instrumenta-
tion in the treatment
of fractures of the thoracolumbar
spine. Spine
1982;7:312-317
4.
McCormick P.Dorsal distraction
and neutral segmental fixation
of
the thoracic
and lumbar
spine: Harrington
and Luque techniques.
are avoided by strict attention to detail when passing the
wire under the lamina. Trimming any sharp surfaces of the
wire will minimize the incidence of a dural breach.
Lossof initial spinal height may occur as the wires that
are affixed to the lamina slide on the rods with bone subsi-
dence. This may be, in part, eliminated by cross-fixating the
rods together, thereby creating a single rigid, quadrilateral
fixation unit.7 Some surgeons prefer not to use the Luque
system for vertebral fractures because of the loss of spinal
height due to excessivebone graft subsidence.
.
Conclusions
Luque rod instrumentation
is a biomechanically neutral
spinal fixation device used for segmental spinal fixation and
kyphotic deformity correction. Its major disadvantage is
that it cannot provide distractive or compressive forces to
the spine like the newer universal spinal systems employ-
ing hooks and screws. Advantages are that the sublaminar
wire fixation technique is technically easy to perform. mul-
tiple points of fixation minimize instrumentation
failure,
and the device has a low profile. Complications specific to
this technique are nerve injury from sublaminar wire pas-
sage and loss of spinal height due to excessive bone graft
subsidence.
In: Benzel
E, ed. Spinal
Instrumentation.
Chicago:
AANS; 1994:
125-144
5.
Hollowell JP,Maiman OJ.Application
of Luque rods and Luque rectan-
gles for stabilization
of the thoracolumbar
spine. In: Fessler RG,Haid
RW, eds. Current
Techniques
in Spinal
Instrumentation.
New York:
McGraw-Hill; 1996:257-271
6.
Drummond
OS, Keene JS.Spinous
process
segmental
spinal
instru-
mentation.
Orthopedics
1988;11 :1403-1410
7. BenzelE.Luque rod segmental spinal instrumentation.
AANS Neuro-
surgical Operative Atlas 1991; 1:433-438

98
Placement of Morphine or Baclofen Pumps
RichardD.Penn
.
Indications
.
Contraindications
.
Advantages
.
Disadvantages
.
The Drug Pump and Catheter System
Direct delivery of medication to a target in the central
nervous system is likely to become increasingly important
with advances in genetic engineering and device design.
Direct delivery avoids the blood-brain
barrier and pro-
vides a high concentration
of medication
where it is
needed and low amounts of medication elsewhere in the
body. Using continuous infusion by drug pump, the med-
ication can be maintained
at the needed level. These
concepts are best illustrated
by the use of intrathecal
baclofen for spasticity
and morphine
for pain,l-3 and
soon other
applications
of this
technology
will be
used for treatment
of brain tumors and the delivery of
neurotrophic
factors
for
degenerative
neurological
diseases.4
The success of the treatment
rests upon the uninter-
rupted delivery of the medication, much as the success of
shunt surgery for hydrocephalus requires a constantly func-
tioning shunt system. The goal of a surgical implant of a
pump and catheter system must be a very high rate of relia-
bility.Assimple as the implant procedure seems to be, sur-
geons must do their utmost to minimize failure of this
expensive technology. In this description of the implant
technique, Iwill emphasize those "tricks of the trade" that
have allowed me to slowlyreduce the number of postopera-
tive problems. Other surgeons may have varying approaches
that also work well,just as there is no single perfect way to
place a shunt.
.
Indications
. Intractablepain
. Failedbacksurgerysyndrome
. Arachnoiditis
704
.
Surgical Technique
Patient Positioning and Drug Pump Placement
Catheter Introduction and Placement
Securing the Catheter in the LumbarPocket
Preparing the Pump for Implantation
Placement of the Catheter Tip
.
Conclusion
. Spacticity
.
Failed oral analgesics or high dose requirement
.
Contraindications
. Activeinfection
. Sulfateor baclofenallergy
. Undesirableside effects (hypertension,nausea,vomit-
ing)
. Historyofnarcoticabuse(relative)
.
Advantages
.
Direct drug delivery
.
Improved quality of life
.
Good pain management
.
Fewsystemic side effects
.
Direct delivery avoids the blood-brain barrier
. Provides high concentration of medication where it is
needed and low amounts of medication elsewhere in
the body
.
Applicable for different medications
.
Disadvantages
. Need to refill pump
. Cost

98
Placementof Morphine or BaclofenPumps
705
. Potentialrespiratory compromise
. Mechanicalfailure of system
. Requiresimplant with potential for breaking/blocking
. Needforbattery replacement
. Rarecomplication is the development of a fibrous mass
at the catheter tip
. The Drug Pump and Catheter System
Thetechnologyusedin implantabledrugpumpsis over20
yearsold.The first drug pump implanted in humans used a
metalbellows reservoir surrounded by a gas that, at body
temperature,exerts a pressure on the bellows system, and,
therefore,on the fluid placed into the reservoir.sToslow the
flowof fluid exiting from the reservoir a long, small-bore
tubeisplacedbetween the reservoir and the exit port of the
pump.The flow resistor slows the rate of flow to 1 to 3 mL
perday,so a 50 mLreservoir will deliver medicine for about
3 weeks before needing to be refilled. This same quite
reliablesystem is still in use today, as well as several other
similardevices. They are simple in design and reliable. The
drawbackto these constant flow devices is that to change
theamount of medicine delivered, the concentrate of medi-
cinein the reservoir must be changed. Depending upon the
clinicalapplication, this mayor may not be a problem.
The second type of device, which has been in use for
over23 years, uses pacemaker technology and a motor-
drivenperistaltic pump to deliver medication. An external
programmer using a radio frequency link instructs a chip
inthe pump on the rate of delivery of medication, and an
appropriate electrical signal is sent to activate the pump.
Thismeans that a highly concentrated
medication
put
into the reservoir can be given at whatever dosage is
needed and the dosage can be easily adjusted to the pa-
tient's needs without a refill. One would expect such a
complex electrical and mechanical system to fail more
frequently than the simple constant flow pump. In prac-
tice they both appear to be highly reliable. The major
drawback to the programmable
drug pump is that its
battery runs down in 4 to 5 years and then the whole
pumphas to be replaced.
The drug delivery is most frequently disrupted by prob-
lems with the catheter going from the pump into the lum-
bar subarachnoid space. Ideally, the catheter
should be
small and soft so that it does not injure nerve roots. Yetit
should be radiopaque, solid, tear-proof and not cause reac-
tive tissue. The ideal catheter has yet to be devised, and
the compromise is far from perfect. The most important
steps in the surgical procedure are those that anchor the
catheter in place so that it does not dislodge, kink, or tear.
Needless to say, damaging the catheter when it is placed
or not putting it into the subarachnoid space brings imme-
diate failure. The Touhey needle for introducing
the
catheter was designed to assess the epidural space not the
subarachnoid
space. It indents the dura before it cuts
through it, and its sharp edge can cut the catheter if the
catheter is introduced and then pulled back. Beware of the
possibility of injury to the catheter if you manipulate it
back and forth out of the needle. Ifthere is any question of
injury to the soft Silastic, take it out and check for a tear
or cut.
The programmable
pump catheters
and connectors
shown in Fig. 98-1 are of the system that is currently be-
ing used by most surgeons in the United States. The con-
stant flow devices are similar in size and in the way they
are implanted. Note that the programmable pump can be
secured to the abdominal fascia by suturing the wire loops
on the pump or by covering the pump with a Dacron
pouch and suturing that to the fascia. The advantage of the
pouch is that it creates an excellent pocket for replacing
the pump. However, if it has to be removed due to an in-
fection or if the pump location has to be changed, its
strong adhesions to tissue make it hard to take out. The
loops can avoid the problem, but a suture at the base of
the pump pocket is hard to place and even harder to re-
move when changing the pump. If the pump is not an-
chored it will very likely turn over making refilling more
than difficult.
Typically two catheters are used, one from the pump to
the lumbar region and the second thinner one going into
the subarachnoid space. The connector between the two
has to be sutured with nonabsorbable sutures and a strain
relief sleeve placed over it. This connection point is a place
where the catheter can be kinked, if care is not taken when
it is put in place.

706
Lumbar and Lumbosacral Spine: Posterior Approach
~
~
Figure98-1
Pumpequipment. The programmable pump witha sideport and wireloopsisshownon the top left and the pumpwithoutsideport
placed within a mesh pouch is on the top right. A single or two-part catheter system can be used (middle left) and a metal connector and sleeveis
used for connecting the two catheter ends (middle right). The U-shaped Silastic catheter and the introducing needle are on the bottom.
. SurgicalTechnique
Patient Positioning and Drug Pump Placement
When using a drug pump to deliver medicine into the lum-
bar subarachnoid space, it is most convenient to place it in a
pocket just belowthe costal margin. The left side is most fre-
quently used to leave room for surgical approaches on the
right side. However,in a patient who might need a feeding
tube, the right side isused. Ialmost always use localanesthe-
sia and sedation with propofol.Thiscontrols spasms in spas-
tic patients, makes surgery painless for patients, and pro-
vides a rapid recovery. General anesthetics or long-acting
agents should be avoided because they may combine with
the newly introduced intrathecal drugs and result in sedation
and in the worst case respiratory depression.
The patient is placed in a lateral decubitus position as
illustrated in Fig. 98-2. Pillows are placed between the
legs, and a strap is used to secure the position of the legs
and the top of the shoulders. Most important, the patient
should be held in a true lateral position so that the spinal
puncture will be easier. The (-arm
fluoroscopy unit is
brought into the room and positioned to provide a pos-
teroanterior (PA)view of the lumbar spine. ProphylacticIV
antibiotics are given prior to starting surgery. Afterappro-
priate prepping and draping, a long-acting localanesthetic
is infiltrated into the region of the intended abdominaland
lumbar incisions and also along the path that will be used
to pass the catheter from one incision to the other. Doing
this early markedly decreases pain during the procedure
and afterward.
Catheter Introduction and Placement
The first step is to place the Touhey needle into the lumbar
subarachnoid
space using fluoroscopy; most often the

98
Placement of Morphine or Baclofen Pumps
707
Figure98-2
Patient positioning: The patient ina lateral decubitus
positionwith the (-arm in place for inserting the catheter under fluo-
roscopic
control.
L3-L4interspaceis the easiest (Fig.98-3). Makesure that
thepediclesare aligned and that the spinous process is in
between.Anoblique approach (40-50 degres) starting 2 to
4emoffthe midline is used to avoid placing the catheter
betweenthe spinous processes where mechanical damage
canoccur.The oblique approach is also important so that
the catheter can easily be passed cephalad. Fig. 98-3A
showshowthe approach looks on the fluoroscopy display,
andFig.98-38 shows the needle penetrating
the dura.
Withthe large Touhey needle (15 gauge) and the dura in-
dentedit is often hard to feel a "pop" as the dura is punc-
tured.lftoo steep an angle of approach is used (like a reg-
ularspinal tap), it may be difficult to pass the catheter
upwardbecause the end of the Touhey needle is up against
theposterior ligament (Fig. 98-38). Also remember that
thenotchon the Touhey needle's trochar points in the di-
T
2-3em
..L
A
Figure 98-3
Angle of Touhey needle introduction.
(A)
TheTouhey needle is introduced
obliquely
into the sub-
arachnoid space to avoid the spinous
processes and to
make it easy to pass the catheter
cephalad.
(8) The top
and lower right figures show how a 90 degree angle will
make the catheter
hit the posterior
ligament
and hinder
the catheter pass.The lower left avoids this problem
due
to the oblique angle.
1
rection of the outlet of the needle. The notch should be
placed up so the needle hole is cephalad. However, this
makes penetration of the dura more difficult because of
the slight curve of the needle. For passing into the sub-
arachnoid space the notch can be down and then turned
180 degrees for passing the catheter. Being sure of spinal
landmarks
and using small steps with frequent
fluo-
roscopy, the introduction of the Touhey needle can be
done in one pass. Obviously a large flow of cerebrospinal
fluid ((SF) is expected once in the subarachnoid space be-
cause the needle is quite large. If the needle placement is
accurate and there is good (SF backflow, it should be easy
to pass the catheter, which is stiffened by a guidewire. If
resistance is felt just as the catheter is coming out of the
end of the needle, the Touhey has to be repositioned
slightly by pulling back a few millimeters and angling up-
ward. The catheter tip should be placed in a regionjust be-
low the conus (usually L1-L2) by fluoroscopy. (SF should
flow back once the guidewire is removed. Ifit does not, an
air block or slight kinking of the catheter may be overcome
by an injection of 1 mLof saline. Ifthe placement is not as-
sured. 1 to 2 mLof contrast material should be injected to
make sure the catheter tip is in the subarachnoid space. It
is possible to be fooled by good (SF flow through a hole in
the dura and to advance the catheter by mistake into ei-
ther the epidural or subdural space. Neither position will
work and a quick check at the time of surgery with
contrast material will be rewarded later.
Securing the Catheter in the Lumbar Pocket
The lumbar incision is made with the Touhey needle in
place to protect the catheter. In an obese patient the inci-
sion may have to be made first to permit the relatively short
B

--.
IIIIIIIIII
708
Lumbar and Lumbosacral Spine: Posterior Approach
midline
Touhey needle to advance into the spinal canal (Fig.98-4).
In any case the curvilinear incision is designed to create a
pocket for the catheter holder and connector. Afterthe lum-
bar pocket has been created, but before the Touhey needle
is removed, an absorbable purse string suture is placed
around the needle as it goes through the lumbar fascia.The
needle is then pulled back and the suture tightened. This re-
duces the chance of (SF tracking along the outside of the
catheter to the lumbar pocket and to the pump. Using an
absorbable suture will avoid a kink or erosion of the
catheter at the fascia. The catheter holder is placed caudad
in the pocket away from the incision and sutured to the fas-
cia. Backflowthrough the catheter is checked to make sure
it is fullypatent.
Theabdominal pocket is made with a curvilinear incision
under the costal margin (Fig. 98-5). The pocket should be
dissected down and toward the midline. In a small patient
the relatively large pump can hit the iliac crest unless it is
toward the midline. The pocket has to be large enough so
the edge of the pump is well away from the incision. In a
thin patient or child,the pump edge can erode the skin. The
pump must not be placed more than 1.5in. deep; a deeper
placement in an obese patient will make refilling difficult
and telemetry impossible. Ifthe incision is lateral, as shown
Figure 98-4
Lumbar incision. An off-midlinein-
cision is made with the Touhey needle in placeto
protect
the catheter.
The catheter is secured to
the fascia with a holder, and a purse string ab-
sorbable suture is placed around the catheter as
it leaves the fascia to avoid cerebrospinal fluid
leaks.
in Fig.98-5, it will disrupt innervation to the refillsiteand
make refills more comfortable.
The next step is to use a regular shunt tube passerto
tunnel from the edge of the lumbar incision to the abdomi-
nal pocket. Previously injected long-acting anestheticalong
the path should make this relatively comfortable. Thetwo
catheters are joined as shown in Fig. 98-1 and the sleeve
pulled over the joint to protect it from kinking. Thecon-
nected catheter is then pulled into its subcutaneous space
from the abdominal region. The catheter is placed ontothe
outport of the pump, and the pump put into its pocket.The
extra catheter is placed under the pump, and the pumpis
sutured by nonabsorbable sutures to the fascia in at least
two spots using either the Dacron pouch or wire loops.
Inspect both pockets to make sure the catheter isplaced
in a way that closure will not cause kinking.Antibioticsolu-
tion can be used to flush both pockets and then a two-layer
closure is performed.
Preparing
the Pump for Implantation
The details for fillingthe pump and preparing it forimplan-
tation are well described in the manufacturer's technical
manuals and will not be described here. However,a few

11IIIII
98
Placement of Morphine or Baclofen Pumps
709
pointsneed to be emphasized. The most dangerous poten-
tialproblemwith the programmable pump is overfillingit. If
overfilled,it can in some circumstances
spontaneously
eject
a largebolus giving a sudden high dose of intrathecal
med-
ication.Therefore, the standard
procedure
is to warm the
pump,withdraw fluid and any air in the reservoir, and then
onlyhalffill it with medication (10 mL).The implant sur-
geonisresponsible for making sure this is done correctly. A
newlyavailablepump design avoids this problem and the
pumpcan be completely filled safely.
Another potential problem with any pump is to disregard
the fluid space in the catheter,
which
cannot
be emptied
fromthe reservoir.This can be up to 0.5 mL.Ifthat amount
ofmedication is flushed through the catheter, it could mean
sudden intrathecal
doses high enough
to cause coma. The
catheter should not be flushed into the subarachnoid
space
ifit has medication in it. Likewise, there is no reason to give
medication to the
patient
through
the
catheter
during
surgery.The possibility
of giving too much is too high; it is
difficultto inject accurately
less than 0.5 mL of fluid. Once
the catheter is attached
to the pump a small bolus can be
programmed to fill the catheter.
Remember
that a major
reason for using a pump
is to provide
a slow infusion
of
medication and avoid bolus doses, which
are much more
likelyto cause side effects.
Placement of the Catheter
Tip
Tounderstand why the best location of the catheter tip is
in the lumbar subarachnoid region below the conus, the
distribution of medication must be understood. The only
Figure 98-5
Abdominal
pump placement.
The place-
ment of the pump should be away from the incision, and
the extra catheter
material
should be placed below the
pump. The pump must be secured to the abdominal
fas-
cia by sutures to the wire loops or mesh pouch.
medications that have been proven to be safe and effective
for chronic intrathecal use are morphine and baclofen.They
are both water soluble so they travel with the bulk flow of
(SF along the spinal canal. Studies have shown that, ifgiven
continuously in the lumbar region, the concentration in the
cervical region will be approximately one quarter that of
the lumbar concentration (Fig. 98-6A).6If the rate of infu-
sion is doubled, the concentration along the cord will also
double.
Introducing medicine in the lumbar subarachnoid space
allows it to mix well with the (SF so it will be evenly dis-
tributed along the spinal cord. The desired concentration at
any given level can be achieved by simply increasing the
rate of flow into the mixing space. If higher doses are
needed in the cervical area, they can be reached safely by
slowly increasing the rate of flow. Placing the catheter
tip along the spinal cord will result in a very high local
concentration and not as even distribution along the cord
(Fig. 98-68). Only in cases where a very localized distribu-
tion is desirable is it necessary to have the tip of the
catheter next to the spinal segment to be treated. This
might be important for delivery of a neurotrophic factor
to an injured segment of the spinal cord. However, most
pain and spasticity patients do not need well localized
medication.
Another reason to have the tip of the catheter away from
the cord is that there is less danger of injury in passing it
and, once placed, having it erode into cord tissue. A very
rare complication is the development
of a fibrous mass
at the catheter tip.7 Such a mass is more likely to create
damage if it isat the cord level.

r
710
Lumbar and Lumbosacral Spine: Posterior Approach
-"
0\ ~
~
.~ "'~~0C::~:.~-;ti"::'/
2
'
~
..
- - . ...~ -
'
....~. .',r
#
.00
_
_____""...; -'
'
.'
-,
~
~..._..
TL&.
'2 t 2
_ _____
-~
." ....,.
----
CONTINUOUS
2
1
1 hour
___ 2 hours
_
4hours
Water Soluble
o
A
C
L
CONTINUOUS
2
1
1 hour
..
2 hours
_4 hours
Water Soluble
o I,;.--
:-.~~
B
tTip in low cervical area
L
Figure98-6
Drug distribution.
(A) The distribution
of drug 1.2. and 4 hours after infusion from a catheter
tip in the preferred
lumbar placement.
Note the concentration
of medication
decreases in the cephalad direction.
with the highest steady state concentration
in the lumbar area. (8)The
distribution
will be symmetrical
around the catheter tip when it is placed above T12. The cervical placement
shown in the figure may introduce too
high a concentration
of drug to the brain stem.

. Conclusion
98
Placementof Morphine or BaclofenPumps
711
Theplacement of a drug pump and catheter allows clini-
ciansto give medication where it will be most effective.
Tobe accepted by nonsurgeons
who will be using it
fortheir patients. the surgical complications have to be
References
1. Penn RD. Savoy SM. Coreas D. et al. Intrathecal
baclofen
for severe
spinal spasticity. N EnglJ Med 1989:320:1517-1521
1
PaiceJA.Penn RD.Shott S.lntraspinal morphine for chronic pain: a ret-
rospective. multicenter
study. J Pain Symptom
Manage 1996: 11:71-80
3. Winkelmuller
M. Winkel muller
W. Long-term
effects of contiguous
intrathecal
opioid treatment
in chronic
pain. J Neurosurg
1996:85:
458-467
4. Penn RD.Kroin JS. York MM. et al. Intrathecal
ciliary neurotrophic
de-
livery for treatment
of amyotrophic
lateral
sclerosis
(phase
I trial).
Neurosurgery 1997;40:94-100
minimized and the system has to work for many years.
The procedure can be done without hospitalization
for
most patients. With some improvements
in pump size
and durability of the catheter the technology should be-
come as acceptable as an Ommaya reservoir or Hickman
catheter placement.
5.
Blackshear
P. Implantable
drug-delivery
systems.
Sci Am 1979:241:
66-73
6.
Kroin JS. Ali A. York M. Penn RD.The distribution
of medication
along
the
spinal
canal
after
chronic
intrathecal
administration.
Neuro-
surgery
1993:33:226-230
7.
North RB.Cutchis PN. EpsteinJA.
et al. Spinal cord compression
com-
plicating
subarachnoid
infusion
of morphine:
case report and labora-
tory experience.
Neurosurgery
1991 ;29:778-784

99
Repair of Myelomeningoceles
Trent L.Tredway
.
Indications
.
Contra indications
.
Advantages
.
Disadvantages
.
Objective
.
PreoperativePlanning
.
SurgicalTechnique
Anesthesia and Positioning
Closureof the Myelomeningocele
.
PostoperativeCare
Myelomeningoceleisone of a constellation of neural tube de-
fects
that
affects
the
entire
nervous
system.
The
myelomeningocele defect results from a failure of primary
neuralation at day26 to 30 during embryogenesis.Thesedates
correspond to the closure ofthe caudal neuropore. The spinal
cord failsto fuse dorsally,leaving a flat plate of neural tissue
referredto as the neural placode.The neural placode also con-
sistsof dysplastic skin at the edges and a central groove that
represents the remnant of the central canalof the spinal cord.
Myelomeningoceles are more commonly located in the
caudalthoracolumbar spine (85%).However,-10% are found
in the thoracic spine and 5%are located in the cervical re-
gion.Abnormalitiesof the spinal canal are frequently associ-
ated with myelomeningoceles, including absence of spinous
processes and laminas, increased interpedicular distances.
decreased pedicleheight, enlarged transverse processes, and
small vertebral bodies.! Furthermore. kyphotic deformities
may be present at birth and scoliosis may present later in
lifecausing backpainand neurologicaldeficits.
Allpatients with myelomeningoceles have an associated
Chiari11hindbrain malformation. Accordingto McLoneet ai,
the morbidity and mortality in myelomeningocele infants
can be attributed to the symptoms arising from the Chiari11
malformation.2
These symptoms
may include
apnea,
cyanosis. bradycardia, dysphagia, nystagmus stridor. vocal
cord paralysis. hypotonia. spasticity,and opisthotonos. Most
of the patients (90%)will have hydrocephalus that requires
shunting. Shunting may be performed before, during, or
after the myelomeningocele repair.
712
l
Although
the prevalence
of myelomeningoceles
has de-
clined over the past several years secondary to prenatal folate
supplementation
and termination
of pregnancy after prenatal
diagnosis. repair of a myelomeningocele
is a procedure that
should be part of the neurosurgeon's
surgical armamentarium.
.
Indications
.
Presence of a myelomeningocele in an infant that isnot
chronically ill or comatose
.
Contraindications
The ChildAbuse Prevention Act3mandates that allpatients
with myelomeningoceles are repaired unless:
. The infant is chronically and irreversibly comatose
. Thetreatment wouldmerelyprolongdying
. The treatment
would be virtually futile in terms ofthe sur-
vival of the infant and inhumane under such circumstances
.
Advantages
. Eliminatecerebrospinalfluid(CSF)
leakage
. Preventinfection

. Disadvantages
99
RepairofMyelomeningoceles
713
. Treatmentcould potentially prolong death if child is
terminally ill.
. Objective
Toobtainclosure of the myelomeningocele and eliminate
(SFleakage, prevent infection, and preserve all neural
function
. Preoperative Planning
Priortosurgicalintervention, the myelomeningocele is cov-
ered with a sterile saline-moistened
Telfa pad and the
infantisstarted on prophylactic intravenous antibiotic ther-
apy.The patient is medically optimized for repair that is
performedwithin 48 hours of birth. The infant is also
placedon latex-allergen prophylaxis to minimize the expo-
sureto the latex allergen. Approximately 50%of patients
withmyelomeningoceleswill demonstrate a latex allergy.4.s
Initialassessment of the infant should include a well-
documentedneurological exam with emphasis on motor
andsensory function of the lower extremities. The infant
shouldalso be examined for signs or symptoms of hydro-
cephalus.Approximately15%of infants with myelomeningo-
celeswillrequire immediate shunt placement.
. SurgicalTechnique
Anesthesiaand Positioning
Generalendotracheal anesthesia is utilized for repair of
myelomeningoceles.The patient is positioned supine on a
foamdonut to protect the myelomeningocele. Once intu-
bated,the patient is turned and placed in a prone position
witha transverse roll under the hips and another roll under
thechest(Fig.99- t). The myelomeningocele defect and the
entirebackare prepped and draped. Careis taken to ensure
that the povidone-iodine
solution does not contact the
neuralplacode.
~'-'
'..
-....
..~
~-
-.
Figure99-1
Positioning for repair of myelomeningocele. Note the
chest and hip rolls.
Figure 99-2
Isolation of the neural placode.Initial incisionalong the
circumferential line formed by the junction of the arachnoid and the
skin.
Closure of the Myelomeningocele
The goal of the operative repair is to preserve neural func-
tion and prevent infection by obtaining good skin closure.
The repair is performed with the use of loupe magnifica-
tion. Also,a nerve stimulator may be helpful in identifica-
tion of neural elements. The initial incision follows the
circumferential line formed by the junction of the arach-
noid and the skin that allows isolation of the neural pla-
code (Fig. 99-2). A rostral incision is then performed to
observe the most caudal intact lamina. Next, the lateral
edges of the neural placode are approximated and the pial
surfaces of the placode are sutured together with a 7-0
absorbable suture (Fig. 99-3). Occasionally, the placode is
too wide and repairing the neural tube defect compro-
mises the perfusion of the tissue. Without closure of the
pial layer, there may be an increase in tethering of the
cord later in Iife.6
The intact dura is then identified and the dura is reap-
proximated
in the midline with a 5-0 nonabsorbable
Figure 99-3
Approximationof the lateral edges of the neural pla-
code with 7-0 absorbablesuture.

714
Lumbar and Lumbosacral Spine: Posterior Approach
Skin
'\... ......-
Underm
\
ined
,.,.........
,
,
I
,
I
~
,,1"
,
--
"
-----
....---------
Figure 99-4
The dura isreapproximated in the midlinewith a 5-0
nonabsorbablesuture.
suture to reconstruct
the thecal sac (Fig. 99-4). Occa-
sionally, large defects may require a thoracolumbar
or
bovine pericardial
dural patch. Next, the lumbosacral
fascia is identified,
incised bilaterally,
and dissected
from
the
sacrospinalis
muscle.
The
fascial
edges
are then reapproximated
with a 5-0 absorbable suture
(Fig. 99-5).
The final
step
in the
closure
of the
myelomeningocele
is reapproximation
of the subcuta-
neous tissue and skin. The subcutaneous
tissue is reap-
proximated if possible and closed with a 5-0 interrupted
absorbable
suture. The skin is undermined
bilaterally
and then reapproximated
in the midline with a 5-0 run-
ning nylon suture (Fig. 99-6).
.
Postoperative Care
The patient is kept in a prone position to relieve pressure on
the wound as well as to keep the wound clean. A sterile
drape is positioned between the surgical site and the anus
to prevent soilage of the wound. Frequent wound examina-
tion and observation for signs of hydrocephalus (increased
References
1.
Cohen AR.Robinson S. Early management
of myelomeningocele.
In:
McClone DG. ed. Pediatric
Neurosurgery.
Philadelphia:
WB Saunders:
2001 :241-259
2.
MeLone DG. Dias L.Kaplan WE. et al. Concepts
in the management
of
spina bifida. In: Marlin AE. ed. Concepts
in Pediatric
Neurosurgery.
Basel. Switzerland:
Karger; 1985:97-106
3.
Federal Register part VI. Department
of Health and Human Services.
1985;50:14878-14901.
Figure
99-5
The fascial edges are approximated with a 5-0 ab.
sorbablesuture.
Figure 99-6
Theskin isundermined and reapproximatedinthemid-
line with a 5-0 nylon suture.
head circumference, tense fontanelle, splaying ofsutures)
are pertinent in the postoperative period. Most infantswill
require a shunting procedure within the first 6 weeksofHfe.
4. PittmanT.KiburzJ.Gabriel K.et al. Latexallergy in childrenwith
spina bifida. Pediatr Neurosurg 1995:22:96-100
5.
Epidemiologic
notes and reports anaphylactic
reactions during gen-
eral anesthesia amons pediatric
patients. MMWR Morb Mort Wkly
Rep 1991:40:437-443
6. ParkTS.Myelomeningocele.In: Albright L.PollackIF.Adelson PD.eds.
Principles and Practice of Pediatric Neurosurgery. New York: Thieme;
1999:291-320

100
Excisionof a Spinal Congenital Dermal
SinuslDermoid
J. GordonMcComb
. Advantages
. Disadvantages
. PreoperativePlanning
. SurgicalTechnique
Positioning
Incision
. PostoperativeCare
. SacralDimples
Acongenitaldermal sinus (CDS)is a tract lined by stratified
squamousepithelium found on, or very near, the midline,
anywherealong the spine, with the overwhelming majority
beinglocatedin the lumbosacralarea.Theselesionsare a
formofaclosedneuraltube defect(NTD)and appearto re-
flecta failure of normal separation between the ectoderm
destined to form the neural tube from that forming the
overlyingskin, with dermal elements becoming trapped
withinor near the developing neural tube. Depending upon
thedegree of incomplete separation, the tract may end in
thesubcutaneous tissue or extend any distance inward to
itsultimate embryological level, which would be the conus
medullarisfor those lesions located in the lumbosacral re-
gionorthe centralcanalofthe spinalcordforthose higher
on the neural axis. As is true with other forms of closed
NTD,a CDSfrequently,though not invariably,has one or
more associated cutaneous markers. A CDS can also be
foundwith other forms of a closed NTD.
ACDShas both dermal and epidermal elements because
theprimitive ectoderm has the capacity to form all compo-
nentsofthe skin.The CDScan peter out into a connective tis-
sueband or enlarge to form one more dermoid cyst. Keratin
debrisfrom the desquamating epithelium and secretions
fromthe sebaceous and sweat glands can produce a slowly
expandingencapsulated mass, containing material that can
varyfrom pearl white and waxy to yellowish and buttery in
consistency.The capsule of the dermoid is usually quite ad-
herent to the surrounding neural structures and becomes
evenmore densely attached in the face of inflammation from
infectionor from the fatty acids of the degenerating material
within the capsules should they gain access to the subarach-
noid space (SAS).Although a dermoid cyst can be found at
one or more locations along the tract, they more commonly
occur at the terminus
and thus usually in the conus
medullaris. Because the abnormal connection between the
dermal ectoderm and neuro-ectoderm during embryological
development is small,the disturbance of the mesoderm con-
densing about the tract is minimal.Asa result, the distortion
of the adjacent supporting structures ofthe spine is insignifi-
cant unless associated with other closed NTD.
.
Advantages
. Repair prevents recurring infections
.
Reduceschronicrisks of meningitis such as hydrocephalus
.
Disadvantages
.
Potential for cerebrospinal fluid (CSF)leak
.
Incomplete excision will recur
.
PreoperativePlanning
Asa result of increasingawareness,more CDSsare being
recognizedearlyin infancyand before becomingsympto-
maticfrominfectionor masseffect.Cutaneousstigmataare
715

716
Lumbar and Lumbosacral Spine: Posterior Approach
very often present and may include angioma, abnormalities
of pigmentation, hypertrichosis, skin tags, and rarely an ad-
ditional tract. Occasionally, another dysraphic condition
such as a lipomatous malformation may accompany the
CDS.The opening of the sinus may be so small as to escape
detection, except with close inspection (Fig. tOO-tA-G).
Hairs may be seen to protrude from the orifice. Debris of
purulent material may drain from the site. The dorsal mid-
line should be very carefully examined in a child who has
one or more episodes of unexplained meningitis.
The depth to which the CDSpenetrates and whether or
not it isassociated with one more dermoid cysts will deter-
mine the clinicalmanifestations (Figs.tOO-2A-C;tOO-3A,B;
tOO-4A-D).The CDSoften goes unnoticed until the patient
presents with infection, either meningitis or an extradural
or intradural abscess (Fig. tOO-SA-F). Either with or with-
out infection, the contents of the dermoid cyst, especially
fatty acids, spilling into the SAS,can produce chemical
meningitis and dense arachnoiditis. The usual hallmarks
of a local infection may be present or completely absent,
even when a deeper-lying
abscess or meningitis exists
(Fig.tOO-6A-C)
Hydrocephalus can develop secondary to increased resis-
tance in the CSFabsorption as a result of one or more bouts
of bacterial or chemical meningitis.
A progressively enlarging dermoid cyst along the tract
will eventually result in compression of adjacent structures,
with the conus medullaris most frequently being affected,
followed by the cauda equina, and only very rarely the
spinal cord above the level of the conus medullaris. Because
there is little, if any, abnormality of the supporting struc-
tures, spinal deformity is not a factor in a patient with CDS,
unless it develops subsequently from a neurological deficit
as a result of infection or neural compression.
Probing the tract or injecting it with a contrast agent is
worthless as a diagnostic procedure and is contraindicated
because doing so may incite inflammation or cause neural
damage. Alumbar puncture is indicated when meningitis is
suspected. Plain x-rays of the spine may show incomplete
fusion of a posterior vertebral arch or may be completely
normal, as is often the case, because a CDSoften enters
through the intraspinous ligament or creates a bony defect
so small that it is not seen. The presence of a dermoid cyst
can widen the spinal canal, as with any other similarly lo-
cated mass. However, there is little need to obtain plain x-
rays of the spine because magnetic resonance imaging
(MRI)is the study of choice. With MRI,it is often possible to
see the tract as it extends into the subcutaneous tissue,
but because of the small size of the tract and the fact that it
can deviate ever so slightly from the midline, there is no
way of guaranteeing how far the tract has penetrated. The
diagnosis of a CDS is clinical and imaging studies are
only augmentative. A negative imaging study does not
exclude the need to explore the tract to its terminus. How-
ever, such studies can be helpful in ascertaining if the CDS
is associated with one or more dermoid cysts or other
closed NTD.
In the presence of meningitis it is best to delay surgical
intervention until the patient receives appropriate antibi-
otic therapy, assuming that the neurological course is
stable. An abscess, be it subcutaneous, extradural, or in-
tradural, requires drainage in addition to antibiotic therapy
and is a surgicalemergencyif the patient exhibitsrapid
progressive loss of neurological function. In the presenceof
an abscess it may be advisable to limit the procedure
to drainage of the purulent material, the removal of the
tract or cyst capsule being delayed until the infectionis
cleared.
. SurgicalTechnique
Positioning
The patient is placed prone and padded to keep the pres-
sure off the abdomen, as for any operation in the lum-
bosacral region. The patient is paralyzed for inductionof
general anesthesia but not thereafter, allowingone theabil-
ity to stimulate suspected nerve roots as needed. Good
lighting and magnified vision, using either higher-powered
Ioupes or the operating microscope, is standard. It hasbeen
our practice to initiate antibiotics and corticosteroidsatthe
beginning of the procedure.
Incision
An elliptical incision is made about the sinus as it emerges
on the skin surface. The skin incision is extended rostrallyin
stages as needed for exposure. It is unusual for the tractto
end before penetrating the deep fascia. Ifthe deep fasciais
completely intact, exploration could be stopped at this
point. In the vast majority of cases, however, one canseea
defect in the deep fascia through which the tract oftenpen-
etrates, often involving the intraspinous ligament between
intact adjacent spinous processes. The deep fascia is incised
in the midline beginning just below the tract and extending
superiorly in steps as needed for visualization. Rongeursare
used to remove the spinous process and laminae rostralto
the tract. The tract is followed into the spinal canal toits
attachment to the underlying dura mater. The dura materis
opened just above the site of penetration by the dermal
sinus, allowing direct visualization as it enters the SAS.
In those cases in which there is no intradural extension
of the tract, the area is very carefully explored to makecer-
tain that there is no small innocuous band that is over-
lookedthat leadsto a more rostrallylocateddermoidcyst.
The patient should have had an MRIstudy preoperativelyto
exclude the presence of such a dermal inclusion cystatone
or morelocationsbelowor in the conusmedullaris.Should
the tract penetrate the dura mater, which is usuallythe case,
additional bone isremovedrostrally.Oneneeds to be prepared
to extend the surgical exposure to the level of the conus
medullaris. In addition to removing the tract, any dermoids
are alsoremoved fromthe cauda equina.
If a dermoid
is found
imbedded
into the conus
medullaris it is meticulously dissected from the surround-
ing spinal cord using standard micromanipulative tech-
niques with the operating microscope. Nerve stimulationis
usedas neededto identifyneuralstructures.Scarringfrom
inflammation, be it bacterial or chemical, can make the re-
moval of the dermoid quite a challenge. Ifthe scarringisof

100
Excisionof a SpinalCongenitalDermal Sinus/Dermoid
717
A-(
D-F
G
Figure 100-1
(A) Typical midline
lumbosacral
location
of a congenital
dermal sinus. (8) A
dermal
sinus surrounded
by an area of capillary
telangiectasia
and a mild hypertrichosis.
(C) The opening
of a tract that has just a very small area of capillary
telangiectasia.
There is
an altered pattern to the hair in the region. (D) A sinus opening
is surrounded
by an area of
hyperpigmentation
and an extensive capillary hemangioma.
(E) Two separate tracts that are
surrounded
by areas of hyperpigmentation.
Only one tract extended
intradurally.
(F) A sinus
opening
with
no cutaneous
stigmata.
(G) A tract
associated
with
a subcutaneous
lipoma
that extended
through
a defect
in the fascia and posterior
neural arches. The lipomatous
malformation
involved the conus that was low-lying
and tethered.
A capillary
telangiectasia
of the skinisalsopresent.

718
Lumbarand LumbosacralSpine:PosteriorApproach
B
c
Figure100-2
(A) The tract extending
to the dura mater. Even though
the tract may have no observable
intradural
component,
it can still be apor-
tal of entry for infection
and therefore
needs to be excised. (B) Intraoperative
photograph
showing
the tract going to the dura mater. (e) Thedura
mater has been opened. One can appreciate
that the tract stops at the dura mater. No intradural
abnormalities
were present.

100
Excisionof a SpinalCongenitalDermalSinus/Dermoid
719
- --
---
A
1/
'>fi,"''::..;;,.$~.~
~- j£'
" B
Figure100-3
(A) The tract that extends to the conus medullaris. No dermal inclusion cyst is present. (B) Intraoperative
photograph
showing the
tractextendingto the conus medullaris.
Figure100-4
(A) Atract that expands into a dermoid cyst at the conus medullaris. (B) Atract that expands into two dermoid cysts. The tract be-
tween two cysts can be very small. (e) Intraoperative
photograph
showing the presence of two dermoid cysts with the one to the right being in the
cauda equina and connected
byjust a small tract to the second dermoid, which is embedded
in the conus medullaris. (D) Intraoperative photograph
showing the sinus tract, which is quite large, extending all the way to the conus medullaris.

720
Lumbarand LumbosacralSpine:PosteriorApproach
A
c
E r-"'.~'I'''
-.: "Ia...
I
_ -~
...1~
~
>:a~.,...r':"~~
Iii
~
~
F
Figure 100-5
(A) A 14-month-old
with a thoracic
congenital
dermal sinus. This lesion went unnoticed
until it became infected.
(8) The opening of
the tract. The skin is red primarily
because of abscess formation.
(C) Tl-weighted
sagittal
magnetic
resonance
imaging
(MRI) study showing the
tract and abscess extending
into the thoracic
spinal cord. (D) Tl-weighted
axial MRI study showing the tract going into the spinal canal. (E) Intraop-
erative photograph
showing the extent of the tract, which was quite large. (F) The spinal cord following
the excision of the tract.

r
100
Excisionof a SpinalCongenital Dermal Sinus/Dermoid
721
A
(
suchintensitythat it isnot possibleto separatethe capsule
fromtheconus,it is best to completelyempty the capsule
ofits contents and then apply bipolar coagulation to the
wallofthe cystinsufficientquantityto destroythe epithe-
liallining. Laser energy has also been reported to be suc-
cessfulin this regard. The SASis thoroughly irrigated with
lactedRinger'ssolution to remove any debris that could con-
tributeto postoperative inflammatory response. The dura
materis closed with a fine running absorbable suture. The
anesthesiologist is asked to Valsava the patient to make
certainthat there is no CSFleak. Additional interrupted su-
turesare placed, if needed, until all CSFleakage is stopped.
Oxidizedcellulose or a similar material is placed over the
exposeddura mater. The deep fascia, subcutaneous, and
cutaneous layers are closed with running absorbable su-
turesofa progressively smaller size.
The rare CDS higher on the spine does not have the
obliquepath of those in the lumbosacral region but extends
intothe canal with only slight rostral angulation through
the dura mater and into the spinal cord. The tract and/or
dermoidis excised from the spinal cord in a fashion similar
to that for the tract and dermoid found in the conus
medullaris.
B
Figure 100-6
(A) A congenital
dermal sinus (CDS) with
subcuta-
neous abscess. Purulent
drainage
is present on the buttocks
from
the sinus located at the inferior aspect of the subcutaneous
area of
redness
and
swelling.
(B)
Intraoperative
photograph
showing
epidural abscess associated with an infected
CDS. (e) Intraoperative
photograph
showing
extensive
scarring
of the cauda equina
and
conus medullaris
associated with a previous
infection.
Also present
is a dermoid
cyst imbedded
in the conus. The presence of the exten-
sive inflammatory
reaction
and scarring
can make removal
quite
difficult.
.
PostoperativeCare
The patient is kept horizontal for a few days to reduce the
chance of CSFleakage. Depending upon the circumstances
of the operative procedure, antibiotics and corticosteroids
may be either stopped in the immediate postoperative pe-
riod or tapered over a period of days.
Ifone is certain that all of the tract or dermoid tissue has
been removed at the time of surgery, the patient is cured
and there is little need for subsequent imaging studies or,
for that matter, neurosurgical follow-up, assuming that the
patient is perfectly normal.
The most important factor influencing the outcome with
a CDSis total excision of the lesion early in infancy prior to
development of infection or neural compression.
.
SacralDimples
The presence of a dimple in the region of the coccyxis
reported to be present in 2 to 4% of all infants (Fig.
tOO-7A-C).Becausethese lesions result from a different

722
LumbarandLumbosacralSpine:PosteriorApproach
embryologicalprocess than a CDSlocated in the lumbosacral
region or above. the embryological termination of dimples
in the coccygealregion is the fascia beneath, with their ori-
entation being straight or caudal, a fact that can be demon-
strated by stretching the surrounding skin either superiorly
or inferiorly.Under all but exceptional circumstances infec-
tion of the central nervous system is not a consideration
with sacrococcigeal
dimples because they cannot extend to
the SAS.The salient point with these dimples is analogous to
that of property
values in real estate-location,
location, lo-
cation. We do not routinely obtain imaging studies for these
lesions. Because only a very small minority of these lesions
becomes infected at a later date, there is no role for prophy-
lactically excising all of these dimples either.
Figure 100-7
(A) Atract located in the coccygeal region. Atract in this location will end in the fascia and has no chance of extending intothe sub-
arachnoid space. Dimples in this location are fairly common and need not be excised. (8) Atypical dimple in the coccygeal region. No neurocuta-
neous stigmata present. (C) Another example of a dimple in the coccygeal region that is also associated with a smallskintag.

101
Resectionof SacrococcygealTeratoma
SigmundH. Ein
. Indications
. Contraindications
. Advantages
. Disadvantages
. Objective
. SurgicalTechnique
Anesthesia
Positioning
Theposteriorapproachto the sacrococcygealspine in refer-
enceto the pediatric (usually newborn) sacrococcygealter-
atomais the prime way this tumor, benign or malignant,
shouldbe approached for its removal. The main reason is
becausethe tumor arises from the coccyx and more often
thannot is exophytic (external) in its nature and growth.
Thereareother variations of tumor mass and presentation
uptoandincluding completely retroperitoneal, pelvic, and
intra-abdominal and as such a combined abdominal and
posteriorapproach may be necessary; this latter approach
israrelynecessary.
. Indications
. Sacrococcygealtumor, benign or malignant, at any pedi-
atricage
. Contraindications
. Unresectable (usually malignant)
sacrococcygeal
tumor
. Advantages
. Mostdirect approach to sacrococcygeal tumor because
the tumor is mostly exterior, attached to the coccyx and
compressingthe rectum anteriorly
. Easyaccessto lateralmuscularattachments
. Easyreconstructionofresidual muscle complex for closure
Incision
Mobilization of Tumor
Blood Supply and Excision of Coccyx
Presacral Extension
Closure
.
Complications
.
Disadvantages
. Closeto anal opening and potential for contamination
.
Access to feeding blood supply and presacral extension
of the tumor may be difficult
.
Without combined laparotomy, large, mostly intrapelvic
and intra-abdominal
tumors are impossible to remove
intact. Even then. such large (especially malignant) tu-
mors are virtually impossible to excise completely (see
Incision section).
.
Objective
To obtain the easiest and best direct visual and manual access
to the tumor and its compressions anteriorly and attachments
laterally;
its origin at the coccyx posteriorly;
and its blood
supply and presacral extension anterior to the coccyx. To put
the operator in the best anatomical situation to have full, safe
control over all aspects of the removal of the tumor.
.
SurgicalTechnique
Anesthesia
The anesthetic should be administered by a pediatric anes-
thesiologist, just as the excision of a sacrococcygeal ter-
atoma in a newborn, infant, or child should be removed by a
pediatric surgeon.Thepediatric patient must be intubated
and muscle relaxed; it makes no difference if nitrous oxide
is administered, unless it is contemplated that the abdomen
723

724
Lumbar and Lumbosacral Spine: Posterior Approach
will be opened and, if so, it is best to avoid the gaseous
distention of the bowel that inevitably comes with nitrous
oxide. Tothis effect a nasogastric tube, at least during the
operation, will help keep the upper gastrointestinal tract
decompressed of air, gas, and fluid contents. A large-bore
intravenous line must be placed in an upper limb for the
possibility of a blood transfusion, although the chances of
excessive blood loss are not great. The patient must be
typed and grouped and cross-matched for at least a unit of
packed cells and more if the tumor is large. The appropriate
monitoring devices are essential. The necessity of an arter-
ialline is debatable; blood gas determinations during this
potentially lengthy operation may be indicated. It is advis-
able to give prophylactic antibiotic(s) (cefoxitin).
Positioning
The position for the posterior approach for excision of a
sacrococcygeal teratoma is prone and jackknife, with a
small bolster under the shoulders and several larger ones
under the pelvis, which positions the coccyx and tumor up
in the air (Fig. 101-1). With this position the abdomen
should not be touching the operating table and the lumbar
spine should simultaneously not be in a "swayback" posi-
tion. The body parts touching the operating table must
be appropriately padded. Although a Foley catheter may
be of some help to decompress the bladder and monitor
fluids both intraoperatively
and postoperatively, a firm
rectal packing intraoperatively
to identify the anteriorly
Figure
101-1
Prone (jackknife)
position
to excise sacrococcygeal
teratoma.
There should be a small bolster
under the shoulders
and
several larger ones under the pelvis, the latter of which push the coc-
cyx and tumor
up in the air. The abdomen
should not be touching
the
operating
table and the lumbar spine should not be in a "swayback"
position.
Note the more commonly
used transverse (chevron)
incision
with the apex 1 or 2 cm above the coccyx. This allows much more ex-
posure than the alternative
up and down midline
incision, which may
be more acceptable
for small tumors.
compressed
rectum
is essential.
The packing can be
soaked with some antiseptic and/or bacterostatic/cidal
compound to minimize rectal or skin contamination. The
rectum can also be irrigated with a prep solution priorto
the packing being inserted and before the operative fieldis
painted.
Incision
The incision of choice is a transverse chevron incisionwith
the apex 1 to 2 cmabovethe coccyx(Fig.101-1).Thisal-
lows much more exposure than the alternative up and
down midline incision.If the tumor, which can be small
(Fig. 101-1), medium (Fig. 101-2), or large (Fig.101-3),is
asymmetric, the side of greatest involvement (extension)
will receive the longer part of the chevron incision.Thein-
cision is invariably mostly above (cephalad to) the tumor,
especially in the large tumors. The key to placing the inci-
sion is to palpate the coccyx, which can always be done,
regardlessofthe tumor size.Ifthe tumor is quitelarge,an
appropriate ellipse of skin can and should be excisedwith
it, but this must be carefully planned and marked intraoper-
atively (Fig. 101-4). If any doubt exists, skin flaps canbe
raised and the appropriate amount of skin excisedat clo-
sure time; the flaps may become accordingly duskyandde-
marcated by that time. Cutting cautery is best used forthe
skin incision because the blood loss with the scalpelblade
on large tumors can be quite significant.
Asthe presacral(retrorectal)tumor increasesinsizeand
rises out of the bony pelvis and up into the retroperitoneal
space, it may require a lower abdominal incision to mobi-
lize the superior portion of the tumor before it can be safely
and easily excised, usually from below. Whether this
Figure 101-2
Medium-sizesacrococcygealteratoma.

_e__
101 ResectionofSacrococcygealTeratoma
725
- fJ- -
abdominalapproach is done first, second, combined, or
staged,and whether the infant or child is to be moved or in
oneposition(either lithotomy or lateral), depends upon the
surgeon's preference.
Suffice it to say that
this ab-
dominoperinealapproach is only successful for benign tu-
morsthat do not invade the surrounding tissues but only
compressthem, and then only rarely so.
Mobilization of Tumor
Ifthe tumor is filled with cerebrospinal fluid ((SF) from
the tumor's choroid plexus (not connected
with
the
patient'scirculating (SF) it is best decompressed to aid in
itsmobilization and removal. Whether the tumor is large
(benign)or small (benign or malignant postchemotherapy)
r
,
Figure 101-4
Ifthe sacrococcygeal teratoma
is quite large, an appro-
priateellipse of skin can and should be excised with it but must first be
carefullyplanned and marked intraoperatively.
Ifany doubt exists, skin
flapscanbe raisedandthe appropriateamount ofskinexcisedat closure.
Figure 101-3
Largesacrococcygealteratoma.
the dissection is best done with a small right-angled clamp
and the coagulating cautery staying just wide of the tumor.
The wound can be kept open quite easily with self-retaining
laminectomy retractor(s) (Fig. 101-5). The larger the benign
tumor is, the more it compresses the tissues (muscles and
fat laterally, rectum anteriorly, peritoneum superiorly), as
opposed to the large unresectable malignant tumor, which
invades rather than compresses the surrounding tissues. It
is difficult to recognize any remnants of the levator sling
:JI'O~
't:tJ'''
Figure 101-5
The wound can be kept open quite easily with self-
retaining laminectomy rectractor(s). If the tumor is asymmetric, the
sideof greatest involvement (extension) will receivethe longer part of
the chevron incision. The largerthe benign tumor is,the more it com-
pressesthe musclesand fat laterally,rectum anteriorly, andperitoneum
superiorly, as opposed to the large unresectable malignant tumor,
which invades,rather than compresses,the surrounding tissues.

726
Lumbar and Lumbosacral Spine: Posterior Approach
(puborectalis,pubococcygeusmuscles) and muscle complex
during this dissection. The least damage to these muscles of
continence will be done by staying close to the tumor as op-
posed to going too wide. Few, if any, vessels require tying
during this dissection but will certainly have to be ade-
quately cauterized. It is best to proceed by this blunt dissec-
tion technique as opposed to the bloodier sharp dissection,
which tends to obscure the planes. It is wise to frequently
finger palpate and dissect because the bulk of the tumor
helps to delineate the cleavage plane. The dissection should
proceed in an orderly fashion from the top (cephalad) su-
perficial to the coccyx (but not through it, because the ma-
jor blood supply is deep to it), down the sides of and around
the tumor and from the bottom (distal portion) proceeding
under the tumor (pushing the rectum anteriorly quite eas-
ily)and then upward (cephalad) toward the narrow neck of
the tumor, housing the coccyx and beneath (anterior) the
sacrococcygeal blood supply (Fig. 101-6). Ifa hole is made
in the posterior rectal wall, it should be closed transversely
Figure 101-6
In mobilizing
the tumor,
the dissection
is best done
with a small right-angled
clamp (Negus, Mixter)
and the coagulating
cautery
staying
close to the tumor.
Few vessels require
tying
during
this dissection.
It is best to proceed
by blunt
(finger)
dissection
be-
cause the bulk of the tumor
helps to delineate the cleavage plane. The
dissection
should
proceed from the top (cephalad),
superficial
to the
coccyx, down the sides of and around the tumor,
and from the bottom
(distal
portion)
proceeding
under the tumor
(pushing the rectum
an-
teriorly)
and then upward
(cephalad)
toward
the narrow neck of the
tumor,
housing
the coccyx and beneath
(anterior)
the major
blood
supply. Note the vessel loop around the latter.
and if possiblecoveredwith anyremnant ofmusclecom-
plex. The neck of the tumor should not be transected from
on top until all other attachments to the tumor are freed,so
that if the vessels are not adequately secured at transection,
the tumor will have been removed and the presacral space
can be easily and clearly visualized and the sacral bleeding
vessels controlled. An operating head light is essential for
this part of the operation.
Blood Supply and Excision of Coccyx
Prior to transecting the coccyx from above (rather than
from underneath), near the lateral muscular attachmentsof
the tumor near the coccygeal neck area are divided usinga
combination of right-angled clamp dissection and cautery
alone (Fig.101-7). The surgeon's nonoperating indexfinger
can easily guide the surgeon in the correct plane untilthe
only tumor attachment
is the coccyx and whatever pre-
sacral extension of tissue there is. This presacral extension
can be easily and safely finger dissected free fromunderthe
tumor
neck just
prior to transecting
the coccyx with
the cautery. Ifthe latter is done slowly the sacral vessel(s)
can be clamped and tied. If there is further bleedingfrom
the presacral area after the tumor is removed, this canbe
easily visualized, grasped with long fine forceps, and
cauterized (Fig. 101-6). Seldom is it necessary to applya
topical hemostatic agent (e.g.,Surgicel) to this area because
of continued oozing.
\\
Figure 101-7
The cephalad
neck (top)
of the tumor
should not be
transected
from
on top until all other attachments
to the tumor are
freed, so that if the vessels are not adequately
secured at transection,
the tumor
will have been removed and the presacral space can beeas-
ily and clearly visualized and the bleeding
sacral vessels controlled. Af-
ter the tumor
has been removed,
if any residual
presacral tumor is
seen or felt, it can easily be excised. Note vessel loop around narrow
neck of tumor,
housing
the coccyx and beneath
(anterior)
the sacral
vessels (Rg. 101-6).

PresacralExtension
101
Resectionof SacrococcygealTeratoma
727
Afterthetumorhasbeenremoved,ifanyresidualpresacral
tumorisseen or felt, it can easily be excised; this usually
amountsto a few centimeters (Fig. 101-6). Gentle finger
dissection,long forceps, and cautery are adequate for its
removaL
Althoughit is extremely rare for this presacral extension
toinvolvethe dura or extend intraspinally, the operating
surgeonshould have been aware of this extension by the
appropriatepreoperative imaging techniques (computed to-
mography,magnetic resonance imaging) and of necessity
consultwith a pediatric neurosurgeon preoperatively. The
malignantvariant of this tumor (never in a newborn) occa-
sionallyinvolves this low spinal area by direct extension,
andas such is rarely, if ever, resectable until it has been
greatlyshrunken by chemotherapy. The rare surgical gym-
nasticsthat are required when the benign internal presacral
extensionisgreater than usual (Altman type IIIand IV)are
describedin the Incision section.
Finally,should the presacral and pelvic dissection pro-
ducea hole in the peritoneum, this can be easily seen and
closedwitha fine running suture.
Closure
Thewound closure is best done in an interrupted, trans-
verse,side to side fashion (as the incision was made) with
fairlyheavysynthetic absorbable sutures (Fig. 101-8). It is
justas difficult to see whatever muscle remnants remain
duringthe closure as it was at tumor removal. Toreconsti-
tutethe midline buttock cleft, the external sphincter tissue
aroundthe distal rectum should be first sutured to the most
distalpresacral fascia in the midline. Subsequently, all su-
turesare placed in a symmetrical fashion from each side
ofthismidline stitch laterally; the sutures are best tied after
theyhaveall been placed. Moreover,it is wise to place a
Silasticself-contained suction drain under this transverse
musclesuture line prior to the sutures being tied and exiting
outastabwound lateral to one ofthe corners of the incision.
Thissuction is important to eliminate any fluid collection;
thereis no evidence that it helps decrease the occasional
woundinfection.Skinclosure should not be with a fine sub-
cuticularsuture or adhesive paper strips because this is an
areaofstress and tension; absorbable skin stitches or staples
leftinplacefor 10to 14days are a stronger option.
Selected Readings
AltmanRP,RandolphJG,LillyJR.Sacrococcygealteratoma: American Acad-
emy of Pediatrics,
Surgical
Section,
Surgery
1973. J Pediatr
Surg 1974;9:
389-398
BassJ.Luks F.Yazbeck S. Long-term
follow-up
of sacrococcygeal
teratomas
with emphasis of anorectal
function.
Pediatr Surg Int 1991 ;6: 119-121
BellMJ,Kosloske AM, Martin
L, et a!. Antimicrobial
prophylaxis
in pedi-
atric surgical patients.
Presented
at the meeting
of the Surgical Section of
the American Academy of Pediatrics,
San Francisco. October
1983
BilikR,Shand ling B, Pope M, et a!. Malignant
benign
neonatal
sacrococ-
cygealteratoma.J
Pediatr Surg 1993;28:1158-1160
Donnellan WA. Swenson
O. Benign
and malignant
sacrococcygeal
ter-
atoma. Surgery 1968;64:834-846
Figure 101-8
Wound closure is best done in an interrupted
fashion
(according to the direction of the incision). It is wise to place a Silastic
self-contained
suction drain under the wound closure exiting out a
stab wound lateral to the incision. Skin closure should be with ab-
sorbable skin stitches.
.
Complications
The newborns are best nursed prone for a few days without
diapers and in an isolette to minimize the chances of stool
and urine contaminating the wound; a Foley catheter may
also help. Wound infection is the most common postopera-
tive complication (15-20%).Therefore, it is reasonable to use
prophylactic antibiotics preoperatively and postoperatively
for 48 hours. The infrequent temporary incontinence of
bowel and bladder (10%)is not unreasonable, considering the
major retrorectal,presacral,and lower retroperitoneal dissec-
tion required for removal of these tumors. However, these
immediate postoperative problems do not seem to be long-
term problems. Other early but less common complications
that can occurare a draining sinus (which usually closeswith
conservative local nonoperative treatment) and wound de-
hiscence (most often closingby secondary intention). It is as-
sumed that if a rare rectoperineal fecal fistula develops in
those cases in which the rectal wall or its blood supply was
violated, it should close in a few weeks; if not, the gastroin-
testinal tract must be defunctioned by fasting and total par-
enteral nutrition or with a temporary colostomy.
Ein SH. Sacrococcygeal
teratoma.
In: Pang D. ed. Disorders
of the Pediatric
Spine. New York: Raven; 1995:361-369
Ein SH, Adeyemi SD, Mancer K. Benign sacrococcygeal
teratomas
in infants
and children. A 25 year review. Ann Surg 1980;191 :382-384
Ein SH, Mancer K, Adeyemi SD. Malignant
sacrococcygeal
teratoma-endo-
dermal
sinus, yolk sac tumor-in
infants
and children:
a 32-year
review. J
Pediatr Surg 1985;20:473-477
Feldman
M, Byrne P,Johnson
MA, et a!. Neonatal
sacrococcygeal
teratoma:
multiimaging
modality
assessment.J
Pediatr Surg 1990;25:675-678
Gonzalez-Crussi
F,Winkler
RF. Mirkin DL. Sacrococcygeal
teratomas
in in-
fants and children:
relationship
of histology
and prognosis
in 40 cases.
Arch Pathol Lab Med 1978;102:420-425

728
LumbarandLumbosacralSpine:PosteriorApproach
..
Grosfeld)L, Ballantine TVN. Lowe D, et aL Benign and malignant
teratomas
in children: analysis of85 patients.
Surgery 1976;80:297-305
Gross RE.Clatworthy
HW)r, Meeker IA)r. Sacrococcygeal
teratoma
in in-
fants and children:
a report
of 40 cases. Surg Gynecol
Obstet
1951 ;92:
341-354
Lindahl H. Giant sacrococcygeal
teratoma:
a method
of simple
intraopera-
tive control of hemorrhage.)
Pediatr Surg 1988;23: 1068-1069
Lisco H. Malignant
tumors
developing
in sacrococcygeal
teratoma.
Ann
Surg 1942;115:378-389
Malone PS, Spitz L, Kiely EM. et al. The functional
sequelae
of sacrococ-
cygeal teratoma.)
Pediatr Surg 1990;25:679-680
Ravitch MM. Sacrococcygeal
teratoma.
In: Ravitch MM. Welch K). Benson
CD,@tai, eds. Pediatric
Surgery. Vol 2. 3rd ed. Chicago: Year Book Medical;
1979:1118-1127
Shanbhogue
LKR,Bianchi A. Doig CM, et al. Management
of benign sacro-
coccygeal
teratoma:
reducing
mortality
and morbidity.
Pediatr Surg Int
1990:5:41-44
Shanbhogue
LKR, Gough DCS. Jones
PM. Malignant
sacrococcygeal
ter-
atoma:
improved
survival
with chemotherapy.
Pediatr
Surg lnt 1989;4:
202-204
Simpson
AH, Porter
A. Davis
A. et al. Cephalad
sacral
resection
with a com-
bined
extended
ilioinguinal
and
posterior
approach.)
Bone Joint
Surg Am
1995;77-A:405-411
Smith
B. Passaro
E. Clatworthy
HW )r. The vascular
anatomy
of sacrococ-
cygeal teratomas:
its significance
in surgical management.
Surgery 1961;49:
534-539
Wooley
MM. Teratoma.
In: Welch K). Randolph)G.
Ravitch MM, et ai, eds.
Pediatric
Surgery. Voll
4th ed. Chicago: Year Book: 1986:265-276

102
SurgicalManagement of Spinal Dysraphism
DachlingPang
. OpenSpinalDysraphism(Myelomeningoceles)
. Embryology
. Preparationfor Surgery
. SurgicalTechnique
PositioningandSterilePreparation
Openingthe Sac
Handlingthe Neural Placode
DuralClosure
SkinandMyofascialClosure
. PostoperativeCare
. Complications
(FirstPostoperativeWeek)
WoundDehiscence
WoundInfection
CerebrospinalFluidLeakage
. OpenMyelomeningocele
with
SegmentalPlacode
Toanexperiencedperson, this is a very broad subject in-
deed,forthe term spinal dysraphism describes many differ-
entformsof congenital malformations of the neural tube.
Thesurgicalrepair of these malformations varies as widely
asthe morbid anatomy of these lesions. There is as yet no
universallyaccepted classification of spinal dysraphism. The
classificationendorsed in Table 102-1 has less to do with
embryogeneticmechanism, structural characteristics, neu-
rology,or region of involvement within the neuraxis than
.Open MyelomeningoceleandKyphectomy
.
SpinalCordLipomas
AnatomyandClassification
.
Embryogenesis
DorsalandTransitionalLipomas
Terminal Lipoma
.
OperativeTreatment
Electrophysiological
Monitoring
DorsalLipomas
TransitionalLipoma
TerminalLipoma
. Outcome
.
LimitedDorsalMyeloschisis
.
Embryogenesis
.
SurgicalIndicationsandTreatment
with whether the lesion is "open" in its external boundary,
a factor that strongly influences the timing and urgency of
surgery. Eachclass consists of malformations that have radi-
cally dissimilar features. The difference in closure tech-
niques necessitates individual description under the appro-
priate lesion heading. The paradysraphic lesions of split
cord malformation and caudal agenesis are not strictly dys-
raphisms and will not be discussed here. Their surgical
managements have been discussed in detail elsewhere.I-3
729

r
730
Lumbar and Lumbosacral Spine: Posterior Approach
Table 102-1
Classification of Spinal Dysraphism
OPENDYSRAPHISM
Open myelomeningocele with terminal neural placode
Open myelomeningocele with segmental neural placode
CLOSEDDYSRAPHISM
Thickened filum terminalis
Spinalcord lipomas
Dorsal lipoma
Transitional lipoma
Terminal lipoma
Dermoid sinus tract
TRANSITIONAL
FORMS
OF DYSRAPHISM
Limited dorsal myeloschisis
Cervical myelomeningocele
PARADYSRAPHIC
MALFORMATIONS
Splitcordmalformations
TypeI
TypeII
Caudalagenesiswithcaudalspinalcordmalformations
NORMAL
NERULATION
Neural
Grrove
Cutaneous Ectoderm
'"
~
~ ~
*j
Neural
Fold
..
.
'.
~
Notochord~
.@
,
.
Endoderm
ELEVATION OF NEURAL FOLDS
. Open Spinal Dysraphism
(Myelomeningoceles)
An open spinal
dysraphism,
synonymous
with open
myelomeningocele,refers to a cerebrospinal fluid (CSF)
filled, membrane-bound
sac with an unclosed segment of
the neural plate (the neural placode) floating on top. It is
the most severe form of neurulation failure during the first
of two phases of spinal cord development (see below).
.
Embryology
Normal development of the spinal cord begins around post-
ovulatory day (POD)22 to 23, when the neural groovedeep-
ens and the neural folds meet in the dorsal midline to form
the primary neural tube (Figs. 102-1 through 102-4).4The
dorsal midline fusion of the neural folds proceed "zipper-
fashion" both caudally and rostrally from the cervicalseg-
ments somewhere near the sixth (cervical) somite. This
phase of development, called primary neurulation, ends
with the formation of the lower lumbar cord segments op-
posite somite 30, around POD28. As the primitive streak
shortens to almost nothing with elongation of the caudal
},
,,,..
Figure 102-1
Primary neurulation. (Top) Demarcation of
a midline neural groove on the dorsal surface of the neural
plate. (Bottom) Elevation of the neural folds, being pushed
in by the enlarging paraxial mesoderm.
Elongation of the
embryo
and stretching
of the cutaneous
ectoderm
also
tends to pull the neural folds toward each other once eleva.
tion has taken place. Note emergence
of the neural crest
cells.

102
Surgical Management of Spinal Dysraphism
731
Actin-likeFilaments
-'-
\,.
~...&~..
Figure 102-2
Primary neurulation. Intrinsicforces
aiding
neural
folds inbending
are provided
by
actin-like contractile filaments within Lcells in the
lateral portions
of the neural folds. The wedge-
shaped Mcells at the median hinge point (median
floor plate) specify the dorsal direction of the in-
bending.
'I'
Median Floor Plate
(Medium Hinge Point)
IN-BENDING
FloorPlate
(MediumHinge Plate)
Notochord
Roof Plate
Sulcus Limitan
..
/
CONVERGENCE,
RECOGNITION AND FUSION
Figure 102-3
Primary neurulation.
Inbending and conver-
gence of the neural folds lead to their dorsal midline fusion.
Glycoaminoglycan molecules appear to be active in the recog-
nition process
between
the approaching
lips of the neural
folds. Note segregation
of neurons within the primitive neural
tube into the dorsal alar plate and the ventral basal plate. sep-
arated by the sulcus limitan. which stretches
from the conus
to the cephalic ventricles.

732
Lumbar and Lumbosacral Spine: Posterior Approach
CSF
Cells from
Meninx
Primativa
Centrum
Notochord
Figure 102-4
Primaryneurulation. Dorsal migration of cellsfrom
the meninxprimitivasurround the primaryneural tube to form dura.
Noteventral migration ofthe neural crest cellsto form the dorsal root
gangliaand the centralprocessofthe dorsalnerve root.
primary neural tube, the caudal cell mass, a cluster of active
pluripotential primitive stem cells appearing around POD
27 to 28,s begins forming (among other caudal embryonic
tissues) a solid cord of future neural cells called the
medullary cord.s This connects with the primary neural
tube, then undergoes central vacuolization (canalization) to
form a secondary neural canal. This second phase of devel-
opment of the spinal cord, called secondary neurulation,
culminates in the conus portion of the cord from somite 30
downward. Afinal process of retrogressive differentiation of
the coccygeal segments of the medullary cord results in the
filum terminalis.2
Most open myelomeningoceles contain a terminal neural
placode, with no recognizable neural tube caudal to the
flattened neural plate.6 It appears that in most cases, the
complete failure of dorsal folding and fusion of the primary
neural plate inhibits secondary neurulation
so that no
conus is formed and the placode ends abruptly. In some
L
cases, remains of an abnormal secondary medullary cord
can be found in the form of a filumlike band attached tothe
lower margin of the neural placode.
The dorsal (apical) surface of the terminal placode corre-
sponds to what would have been the ependymal (inner)lin-
ing of the cord if neurulation had taken place, and its ven-
tral (basal) surface corresponds to the outer surface ofthe
"would-be" neural tube. The sensory and motor roots from
the neural placode therefore project from its ventral surface
only, the more lateral sensory roots issuing from the alar
plates and the medial motor roots from the basal plates.The
cutaneous ectoderm from each side of the embryo normally
destined to fuse in the midline is kept widely apart bythe
unneurulated placode. The dorsal surface of the placodeis
therefore either "naked," or covered by an epithelial mem-
brane of variable thickness grown in from the surrounding
pia-arachnoid
layer (Fig. 102-5). The placode also effec-
tively prevents dorsomedial migration of the mesenchyme,
and thus the completion of the posterior neural arch(hence
the term spinal biJida), dorsal paraspinous muscles,and
lumbodorsal fascia.
Because the meninges develop adjacent to the basalsur-
face of the neuroepithelium, only the ventral (basal)surface
of the unneurulated
placode receives meningeal invest-
ment.6.7As(SF accumulates between the ventral surfaceof
the placode and underlying leptomeninges, the flatplacode
is subjected to increasing dorsally directed forcesand,lack-
ing dorsal support from integumentary and myofascialtis-
sues,isultimatelypushedout dorsallyto rideonthedome
of the distended cyst (Fig.102-6). The remainingdorsal
wall of the sac on each side of the placode is composedof
leptomeninges that were also ballooned out by the (SFand
stretched between the lateral edge of the placode andthe
margin of abnormal skin. Intact dura lines the ventralpor-
tion of the sac, being also prevented from dorsal midlinefu-
sion, and instead fuses with the margin of the unclosed
skin, dorsal musculature, fascia, and periosteum ofthein-
complete neural arch on both sides of the myelomeningo-
cele sac.6

102
SurgicalManagementof SpinalDysraphism
733
Neural Plate
NeuralCrest \
"
Notochord
t'.
.-/1:.
Sensory Root
Centrum
..."
".
~
Ventral
Dura
Neural Placode
Sensory
Root
Skin
,/
Lu!"bodorsal
Fascia
>U.J
Dorsal
Musculature
Centrum
Figure 102-5
Formation
of an open myelomeningo-
cele. (Top) Complete
failure of elevation
of the neural
folds.
(Middle)
Sensory and motor
roots develop
on
the ventral surface of the flat neural placode; the sen-
sory root
occupies
a more
lateral
position
than the
motor
roots. (Bottom)
The meninx primitiva
cells can
only form the ventral dura. fail to cover the dorsal sur-
face of the neural placode.
and fuse to the cutaneous
neural junction
on each side.
Figure 102-6
Formation
of open myelomeningo-
cele. The increasing
cerebrospinal
fluid pressure on
the ventral
side of the
neural
placode
ultimately
pushes the placode out onto the dome of an enlarg-
ing cyst. Leptomeninges
stretch
between
the lateral
margin of the neural placode to the edges of the ab-
normal
skin. Note ventrally
streaming
nerve roots
from
the ventral
surface of the placode.
The lateral
edges of the dura are fused with
skin. lumbodorsal
fascia. dorsal paraspinous
muscles. and periosteum
of the bifid neural arch.

734
LumbarandLumbosacralSpine:PosteriorApproach
.
.
Preparation for Surgery
The goals of surgical management of open myelomeningo-
cele are (1) preservation of neural tissues, (2) reconstruc-
tion of the dural tube. (3) securing sound myofascial and
skin closure. and (4) minimizing the chances of future
retethering of the cord. Most open lesions are closed within
24 hours after birth. Ifthe child is initially unstable. closure
may be safely delayed for up to 72 hours without an in-
crease in complications. Performing surgery after that time
carties a substantial risk of meningitis. wound abscess.8.g
and neurological deterioration.
The preliminary workup consists of chest and spine films.
routine blood work, and cross-typing for blood products. A
quick neurological assessment suffices to document the ex-
act sensorimotor level of the lesion as well as whether gross
hydrocephalus necessitates simultaneous external ventricu-
lar drainage. About 10 to 15% of patients
with open
myelomeningocele are born with alarming signs of gross
ventriculomegaly
such as severe macrocephaly.
tense
fontanels. and cardiorespiratory
instability. The infant
should also be checked for pulmonary insufficiency and for
coexisting life-threatening anomalies, such as renal agene-
sis and irreparable cardiac defects. normally using ultra-
sound studies. Invariably fatal chromosomal abnormalities
such as trisomy 18 must be verified with an emergency
karyotype. Presence of any such untreatable lesions incom-
patible with a decent quality of life should prompt a realis-
tic discussion with the parents and recommendation of no
intervention.
While awaiting surgery. the infant is placed prone. and
the placode is protected by a warm. sterile, saline-soaked.
nonadherent
dressing. reinforced with a plastic wrap to
minimize rapid desiccation. An intravenous line is started
and antibiotics are given if there is a history of premature
rupture of membranes.
.
Surgical Technique
Positioning
and Sterile
Preparation
The operating room must be kept warm to avoid hypother-
mia during sterile preparation of the operative field. All
prep solutions are warmed to body temperature. The infant
is intubated in the lateral position
or supine with a
doughnut-sponge on the back to accommodate for the sac.
The pelvisis propped up with a horizontal roll to render hy-
perlordosis of the lumbar spine and maximum relaxation of
the skin and softtissues.The membranous portion of the sac.
containing the exposed neural placode. is irrigated profusely
with antibiotic-reinforced saline (Betadine compounds have
been shown to be neurotoxic). and the surrounding skin
scrubbed with the usualantiseptic agents.
Opening the Sac
I prefer using magnification at the very beginning. The im-
portant structures of the neural placode are best identified
by entering the sac through the diaphanous leptomeningeal
membrane approximately halfway between the marginof
healthy skin and the edge of the placode (Fig. 102-7).
Neural tissue of the placode is recognizable by its pink,felty
surface. transverse wrinkles, and a straight. longitudinal
median raphe (Fig. 102-8). Cutting too close to neuraltis-
sue beforeknowingthe locationofthe sensorynerveroots,
being hidden from view. may damage the dorsal rootentry
zone (Fig. 102-9). The epithelialized membrane shouldbe
avascular except where there are large. obvious crossing
veins. so that copious bleeding from the cut edges usually
means breaching neural tissue. Bleedingiscontrolledwitha
pair of low current. fine bipolar cautery forceps equipped
with continuous irrigation.
After the initial gush of CSFand collapse of the cyst.the
edges of the neural placode are gently flipped up to identify
the exact location of the ventral nerve roots and feedingar-
teries. Many crossing blood vessels have to be coagulated.
Thepearlyepithelium must be meticulouslytrimmedcir-
cumferentially from the placode to avoid later occurrenceof
inclusion dermoid cyst. At the caudal extreme ofa terminal
placode, the epithelial membrane may remain thin. orone
may encounter a band like thickening (filum terminalis)
representing a remnant of the secondary medullarycord
that must be divided to free the tip. At the rostral extreme
of the placode. careful incision of the epithelium-neuraltis-
sue junction
on both sides exposes the delicate bevel-
shaped transition between the unneurulated. flat placode
and the neurulated cylindricalspinal cord.Attheapexof
the bevel. the central canal of the normal cord canbeseen
unfurlingintothe medianrapheofthe placode.fromwhich
CSFis sometimes slowing oozing (Fig.102-10).
Handling the Neural Placode
The neural placode is always handled
gently with jeweler's
microforceps.
If the placode is pliable and thin enough. itis
rolled on itself and sewn up with several 8-0 nylon sutures
picking up the delicate pial lining on the placode's ventral
margins
(Fig. 102-11).
There is no evidence
that this im-
proves neurological function. but it. at least intuitively.re-
duces the "sticky" surface from a flat plate to a line.making
the cord less vulnerable to later tethering. Ifthe placodeis
too thick or too stiff, it should not be forced to roll upfor
fear of mechanical and vascular compromise. It is question-
able whether some thin placodes in fact contain neuralele-
ments
"wired"
for voluntary
movements,
but tugging too
hard on the proximal spinal cord will cause ascensionofthe
neurological level. In addition. the caudal end of the placode
must be inspected for the presence of a neurulated cordin
the rare case of a segmental placode, to avoid liberallyhan-
dling what might be assumed to be the "free" end of the
placode.
Paradoxically,
small
and compact
neural
pIacodes are
likely functional and therefore should be treated with
extra care. whereas large. thin. and "spread-out" placodes
are
mostly
found
in mid
to high
thoracic
lesions and
are usually
nonfunctional.
It is sometimes
advisable to
cordectomize these large placodes. for although they donot
convey
volitional
movements.
the
isolated
neuronal
pools within
them can produce
pathologically
hyperactive
local reflexes underlying high-pressured. spastic bladders

CLOSURE OF OPEN PLACODE
102
Surgical Management of Spinal Dysraphism
735
"
Dura
Skin
~~
:.6 ..-
~~
'..
>'.~~;
,,.;
Dorsal
Musculature
~~~~
Bifid lamina
~
;._:.,.-.~\
-' ..
. .
associatedwith ureteral reflux. Reflux is less problematic
whenthebladderis renderedflaccidfoHowingelimination
ofthistype of placode. The now tubular (or compact) pla-
codeis dropped back into the ventral dural sac.
DuralClosure
Theouter margins of the dural flaps are created by sharply
incisingthe ventral dura from the lumbosacral fascia and
periosteum(Fig. 102-7, middle), and a new dural tube is
reconstructed in the midline (Fig. 102-12). One aims to
Figure 102-7
Steps of surgical repair of an open
myelomeningocelewitha terminalplacode.(Top)Excision
ofthe arachnoid dome of the cyst bycutting closeto the
margin of the neural placode and the junction between
membrane and healthy skin. (Middle)Placode has been
rolled up with pia to pia microsutures. The lateral dural
margins are sharply detached from their lateral attach-
ments and closed in the midline over the placode. (Bot-
tom) Closureoftriple-layeredflapsof muscles,fascia,and
skinoverreconstructed duraltube.
obtain as capacious a dural sac as possible commensurate
with the size of the placode, the theory being if the placode
passively flops freely within a large (SF space it is less likely
to adhere to the dorsal dura. This is almost always achiev-
able with the patient's own dura, even if some of the perios-
teum overlying the bifid neural arches may have to be
mobilized with the dura proper to enlarge the sac. In the
rare event of insufficient dura, a substitute has to be used. I
prefer bovine pericardium, not because it is proven less
conducive to retethering, but because its texture is most
compatible with newborn dura and least likely to lead to
suture-hole leakage of (SF.At the end of closure, the suture

736
Lumbarand LumbosacralSpine:PosteriorApproach
Figure 102-8
Open myelomeningocele with a terminal neural
placode. Note longitudinal median raphe on the neural placode.
Small arrowheads outline margin of placode. Glistening lep-
tomeningeal membrane stretches between placode and skin. Large
arrowheads outline skin edge. Triangular end of the sac is toward
the anus.
~;~. ..
r~~~"
.""\..
Figure 102-9
Open myelomeningocelewith terminal placode. Cut-
ting intothe leptomeningealmembrane just outside the marginofthe
placode.
line is tested with a Valsalvamaneuver held at a pressure of
35 to 40 cm of H20for 10seconds.
Skin and Myofascial Closure
The true size of the skin defect is only apparent after excis-
ing the epithelialized membrane back to full-thickness skin
(Fig.102-10). The transitional ring of flimsy epidermis. lack-
ing dermis and the subdermal capillary network, is a poor
medium to sustain tensile sutures. "Healthy bleeding" from
the cut edge of "good" skin should be controlled as much as
possible by pressure and the least amount ofcautery.
Figure 102-10
Open myelomeningocele
with terminal
placode after
leptomeningeal
membrane
has been removed. Trimmed neural placode
fell back on to bottom
of the original sac. Ventral dura (D) lines the bot-
tom of the sac and fuses with the skin margin held by hook. Rostralmar-
gin of the placode (white arrow) is continuous with the normal spinalcord.
Figure 102-11
Open myelomeningocele. Terminal placodebeing
"neurulated" with 8-0 microsutures stitching pia to pia. Notesensory
(dorsal) roots on each side of the placode margin. Note blind endof
the terminal placode on the right of picture, and the beveledspinal
cord-placode transition on the left sideof picture.
Defects of 2 to 3 cm diameter can usually be closed primar-
ily after the subcutaneous
layer is gently lifted and mobilizeda
short distance centrifugally, just enough to reduce the tension
on the skin edges. The subdermal
layer is closed with inter-
rupted absorbable sutures and fine skin sutures. Several surgi-
cal techniques
have been developed
to minimize suture line
tension in large myelomeningocele
defects. Lateral relaxing in-
cisions with bipedicle flap closure in the midline have been er-
fective,10 but the relaxing incisions themselves
then require
skin grafting at the same time or at a later date. Complex mul-
tiple rotation skin flaps have also been tried. but this necessi-
tates extensive skin undermining
and still does not altogether

102
SurgicalManagement of SpinalDysraphism
737
Figure102-12
Open myelomeningocele. Closure of dural tube.
Paraspiniousmuscles and periosteum
of bifid neural arch are exposed
oneachsideofthe closedduraltube. Anusto the rightofpicture.
eliminateall tension spots. I favor using composite skin-
muscle(myocutaneous) flaps for large defects (Fig. 102-13).
Thereare three advantages to this technique. First. cadaver
vaso-Iatexstudies show that a rich anastomosis
exists be-
tweenthe vasculature of the skin overlying the gluteus max-
imusand the latissimus dorsi muscles on each side and across
the midline; the muscles
themselves
are supplied
by the
glutealand thoracodorsal
arteries,
respectively.
As long as
thesearterial pedicles are preserved
and the four myocuta-
neousblocksare mobilized as one unit. the vasculature of the
lumbosacralskin is well maintained
when the four muscles
areapposedin the midline, even if the short paraspinous arte-
rialperforators deep to the muscle are taken.l1 Second. be-
causethere is no undermining
of the subcutaneous
tissues,
theskintension is mostly absorbed by the muscle closure. Fi-
nally,thismethod results in a triple-layered (muscle.subcuta-
neoustissue. and skin) coverage over the dural closure and
addsextrainsurance against CSFleakage. The flaps survival
rateevenforenormous defects isgreater than 92%with this
technique,and the blood loss and extra anesthetive
time are
quite acceptable.
II
Figure102-13
Myocutaneous flaps closure of large skin defect. The
thoracodorsaland gluteal arteries are traced on their respective flaps.
LD,latissimusdorsi muscle; G, gluteal muscle.
It is basic principle to avoid leaving a large amount of blood
or serum under the flaps; a seroma or hematoma
increases
skin tension and prevents "sticking down" of the flap to the
underlying tissue from which neovascularization
for the flap
must be derived.lfthe
flaps are large and "wet," a small drain
without negative suction may be left in for 24 hours. Iam gen-
erally reluctant to use suction drains for fear of perpetuating
a
CSFfistula and would rather rely on meticulous hemostasis.
.
Postoperative Care
The immediate postoperative concern is wound care. The
wound should be kept moist with a generous spread of baci-
tracin ointment and coveredwith a light nonstickingdressing
(e.g.,Telfa),fashioned so that it can be easily lifted up for in-
spection several times daily.The infant isnursed prone at all
times for the first 7 to 10 days,and the hips should be hyper-
extended by a horizontal rollunder the anterior iliaccrests to
allow maximum relaxation of the back skin. Even though
temporary hypothermia may be problematic. a heat lamp is
seldom used because direct radiated heat on the wound may
induce relative ischemia to the flaps from hypermetabolism.
Hydrocephalus normally does not pose a problem until 5
to 7 days after closure of the open sac, which acted as a
pressure reservoir and absorbed much of the CSFpressure.
The occipitofrontal head circumference is measured daily
and the tension of the patent fontanels is checked fre-
quently. It is preferable to await insertion of a CSFshunt un-
til the back wound shows initial healing without signs of
breakdown. CSFleakage, or infection. On the other hand.
high CSFpressure may cause a leak, and often precipitates
early signs of brain stem compression due to the Chiarimal-
formation. Ifthere is any question with the integrity of the
wound and more time is needed. the ventricles can be de-
compressed by serial ventricular taps.
An indwelling bladder catheter is usually left in for as
long as the infant is prone. Intermittent catherization isdif-
ficult in this position. Male infants are recommended to
have circumcision before hospital discharge for ease of
clean catheterization by the parents at home. The infant can
usually be nursed in the lateral and supine positions after 5
to 7 days, depending on the strength of the skin closure. and
at this time urodynamics and renal ultrasound scan are per-
formed to assess intravesicular pressure, bladder capacity,
and ureteralreflux.Cleanintermittentcatheterization(CIC)
is recommended if the leak point pressure on cystometry is
over 20 cm H20,or if there is demonstrable reflux.
.
Complications (First Postoperative Week)
The operative mortality for children undergoing repair of a
myelomeningocele should be close to 0.12-14The most com-
mon cause of postoperative death is related to hindbrain
dysfunction (73%).15.16
but this seldom occurs acutely in the
first week of life.Most of the immediate complications per-
tain to the wound itself.
Wound Dehiscence
A study of the nutritional status of newborn infants who
have had myelomeningocele surgery. using body weight.