Keywords
retrosigmoid craniotomy - retrosigmoid approach - primary dural closure - CSF leak
Introduction
In skull base neurosurgery, the retrosigmoid approach has been widely described and
utilized for access to various pathologies of the posterior fossa and the cerebellopontine
angle.[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8] Despite the widespread utilization of this approach, one significant complication
remains frustratingly common: postoperative cerebrospinal fluid (CSF) leak form the
inner ear, nares, or the surgical site itself with varying incidence (0–22%).[7]
[9]
[10] In a 2004 pooled-group analysis of 2,273 patients from 14 studies, Selesnick et
al. reported an overall CSF leak rate of 10.6% after retrosigmoid surgeries.[11]
Many techniques have been described to reduce rates of postoperative CSF leak after
retrosigmoid craniotomy, including meticulous dural closure in single or multiple
layers with bony reconstruction, the use of varying allografts (fascial, fat, and
or muscle autografts), postoperative lumbar drainage, and pressure or vacuum dressings
with varying degrees of success.[1]
[7]
[10]
[12]
[13]
[14] Two series have reported the incidence of postoperative CSF leak following retrosigmoid
craniotomy without the use of allo- or autograft. In 1999, Gal and Bartels reported
an incidence of 2.9% with only the use of bone wax, and Yamakami et al reported an
incidence of 4% with a similar technique in 2004.[15]
[16]
In this study, we report our experience in utilizing a technique to maintain the integrity
of the dural flap, allowing for primary dural closure from the retrosigmoid approach.
We report the incidence of postoperative CSF otorrhea, rhinorrhea, and incisional
leak or pseudomeningocele. The technical nuances of primary dural closure of the retrosigmoid
approach are described within the manuscript.
Methods
A retrospective chart review was performed to identify the primary surgeon's (L. Madison
Michael) patients who underwent retrosigmoid craniotomy from February 2009 to February
2015. Electronic medical records—including discharge summary, history and physical,
operative and radiographic magnetic resonance imaging (MRI) and computed tomography
(CT) reports, and clinic notes—were reviewed to determine the diagnosis of each patient
and to confirm each surgical approach with particular emphasis on development of CSF
rhinorrhea, otorrhea, pseudomeningocele, or incisional leak by the first surgical
follow-up. Patients were excluded if they did not undergo a retrosigmoid craniotomy.
Individuals performing the chart review were all trained by one individual (Garrett
T. Venable) to ensure consistency of review and coding. To further verify accuracy,
two individuals (Garrett T. Venable and L. Madison Michael) reviewed all patient records.
We collected the following data points for each study patient: (1) age at the time
of surgery, (2) primary diagnosis, (3) sex, (4) time to first post-surgical follow-up,
(5) presence or absence of postoperative CSF leak or pseudomeningocele, (6) length
of hospital stay (LOS), and (7) any other complications of surgery (e.g., meningitis
or wound infection).
Our primary outcome measure was development of a CSF leak (otorrhea, rhinorrhea, or
incisional leak) during the initial hospital stay or by the first surgical follow-up.
The institutional review boards of Methodist Le Bonheur Healthcare and the University
of Tennessee Health Sciences Center approved this study. All statistics were calculated
using SPSS v.22 (IBM, Armonk, New York).
Results
Case Series
Eighty-six patients were identified who underwent retrosigmoid craniotomy between
February 2009 and February 2015. Patient characteristics can be found in [Table 1]. Median age at the time of surgery was 55 years (range, 21–80 years) with 53 (61.6%)
females and 33 (24.8%) males. The most common indications for retrosigmoid craniotomy
were microvascular decompressions (50, 58.1%) and tumor resection (31, 36%). Primary
dural closure was possible in all cases, and no patients required allo- or autografts
to repair the dural defect. Median LOS was 3 days (range, 1–33 days). Median time
to first postoperative follow-up was 24 days (range, 12–679 days), and 5 (5.8%) patients
were lost to follow-up. No patients developed a CSF leak (otorrhea, rhinorrhea, or
incisional leak) or pseudomeningocele by hospital discharge or at the first postoperative
follow-up visit. One (1.3%) patient developed a postoperative wound hematoma that
resolved without surgical intervention. No patients developed a postoperative infection.
Table 1
Characteristics of patients undergoing retrosigmoid craniotomy
|
Variable
|
Value
|
|
Age, years; median, range
|
55, 21–80
|
|
Gender
|
|
|
Male
|
33 (38.4%)
|
|
Female
|
53 (61.6%)
|
|
Indication
|
|
Microvascular decompression
|
50 (58.1%)
|
|
Tumor
|
31 (36%)
|
|
Acoustic neuroma
|
7 (8.1%)
|
|
Metastasis
|
7 (8.1%)
|
|
Meningioma
|
7 (8.1%)
|
|
Glioma
|
3 (3.5%)
|
|
Vestibular schwannoma
|
3 (3.5%)
|
|
Hemangioblastoma
|
3 (3.5%)
|
|
Chordoma
|
1 (1.2%)
|
|
Dermoid cyst
|
2 (2.3%)
|
|
Brainstem cavernous malformation
|
2 (2.3%)
|
|
Cerebellar abscess
|
1 (1.2%)
|
|
Length of stay, days; median (range)
|
3, 1.2–33.1
|
|
Time to first follow-up, days; median, range
|
24, 12–679
|
Surgical Technique
Scalp incisions vary depending on pathology. The incision for microvascular decompression
surgery is linear and is 5 cm in length. It is deepened to expose the bone, and no
attempt is made to develop soft tissue layers. For all other pathologies, the incision
is C-shaped and retroauricular. The galea and skin are then carried forward as the
first layer followed by mobilization of the underlying muscles inferiorly and that
of the musculoperiosteal layer superiorly. In all cases, craniectomies are performed,
and the posterior aspect of the transverse and/or sigmoid sinus is exposed ([Fig. 1]). If mastoid air cells are visualized, bone wax or bone paste is used to seal them
to exclude communication with the middle ear. The dura is opened immediately posterior
and inferior to the sigmoid and transverse venous sinuses, respectively, ([Fig. 2]), and stay sutures are placed on the dural edge to enhance exposure ([Fig. 3]). The posterior fossa dura maintains its position on the moist surface of the cerebellum,
and a wet cottonoid is placed on the top of the dura to prevent drying ([Fig. 4]). Intermittent irrigation of the cottonoid and dura is performed throughout the
case as necessary to avoid desiccation. Primary closure of the dura begins inferiorly
and is carried superiorly using interrupted 4–0 Nurolon (Ethicon) sutures ([Fig. 5]). Complete closure was possible in all patients. DuraSeal (Integra) tissue glue
is then injected over the suture line. A dry piece of Gelfoam (Pfizer) is placed within
the epidural space, and contoured titanium mesh is used to reconstitute the bony defect
([Fig. 6]). No fat, lumbar drain, or head dressing is used. The wound is closed in layers.
The skin is closed in a running fashion using 3–0 Rapide (Ethicon) suture. Dermabond
(Ethicon) is placed overlying the incision as the sole dressing.
Fig. 1 Bony exposure of a typical microvascular decompression case. The distal transverse
sinus and proximal sigmoid sinus are visualized.
Fig. 2 The dura is initially incised along the inferior border of the transverse sinus.
It is then carried inferiorly just posterior to the sigmoid sinus.
Fig. 3 Stay sutures are placed along the venous sinus side of the dural opening.
Fig. 4 The dural flap is left directly on the moist surface of the cerebellum and is covered
by a moist cotton patty.
Fig. 5 Primary closure of a microvascular decompression case is demonstrated here.
Fig. 6 Reconstruction of the bony defect is performed using titanium mesh.
Discussion
Postoperative CSF leak following retrosigmoid craniectomies remains a frustrating
complication and represents a large economic burden to patients and hospital systems.
In a 2004 cost analysis of postoperative CSF leaks and cost effectiveness of dural
sealants, Grotenhuis found that 44 (10.7%) of 412 total patients who experienced postoperative
CSF leaks accrued 21.7% of the total cost for the group and €17,412—or $19,088—more
per procedure.[17] A more recent study by Hendricks et al found the average cost of readmission for
postoperative CSF leak after endoscopic transsphenoidal surgery to be $24,613;[18] additionally, patients with elevated body mass index (BMI) may have an increased
risk for CSF leak.[6] Many techniques addressing the problem of CSF leaks have been described through
the years with varying rates of success, including galeal, fascial, or fat grafts
alone or in combination[7]
[10]
[19]
[20]; meticulous primary closure of anatomical layers[9]
[15]
[16]; dural sealants[21]; dural allografts[22]; bone cement reconstruction[5]
[14]; postoperative lumbar drainage[2]
[19]
[23]
[24]; and postoperative compression dressing.[10]
[20] Two recent studies have reported 0% CSF leak rates. Ling et al describe their closing
technique during retrosigmoid craniectomies, which included autologous fat graft overlying
the primary dural closure (with or without the use of a dural allograft patch), Medpor
Titan cranioplasty, and 1 to 2 days of postoperative pressure dressing. Their mean
LOS was 3.8 days with no reported complications at a median follow-up of 1 year.[10] Eseoneu at al report a 0% CSF leak when using calcium phosphate cement cranioplasty
as well as a statistically significant decrease in CSF leaks as compared with polyethylene
titanium mesh cranioplasty (0% vs 4.5%, p = 0.03).[10]
[14]
Traditional teachings involving the retrosigmoid approach recommend opening the dura
along the periphery of the craniotomy/craniectomy and carrying it toward the venous
sinus.[25] In comparison with the dura adjacent to the venous sinuses, the dura overlying the
cerebellum is quite thin and friable; folding the dura upon itself or retracting the
dural leaflets with stay sutures with this technique can more easily lead to dehydration
and retraction of the delicate dural flap as the case progresses. Primary closure
of the dura, then, can be difficult in this situation, necessitating additional measures
to mitigate postoperative CSF leaks. To ensure primary closure of the dura in cases
involving the retrosigmoid approach, we have found it necessary to shift the dural
incision much closer to the venous sinuses. In all cases, exposure of the posterior
aspect of the venous sinus is performed as it reduces the need for cerebellar retraction.
To ensure that venous sinus injury does not occur during the exposure, we prefer the
use of a craniectomy for better exposure. With direct visualization of the venous
sinus, it is possible to open the dura at the point of maximum thickness. Retraction
sutures are placed through the dural edge on the venous sinus side in an effort to
maximize visualization, reduce cerebellar retraction, and avoid rundown of blood from
the epidural space. Avoiding the use of retraction sutures or folding of the dura
on the cerebellar side allows the layer to remain moist throughout the procedure.
A wet cottonoid prevents the outer layer from becoming dehydrated by the light of
the microscope. Following the completion of the intradural portion of the procedure,
the dura is re-approximated using interrupted sutures. Interrupted sutures—as opposed
to a running suture—are felt to lead to a watertight closure, as supported by the
work of Megyesi et al.[26] DuraSeal is used to reinforce the suture line and is in no way a substitute for
incomplete dural closure. It is biodegradable and prevents the patient from undergoing
a possible second incision to obtain fat autograft. Dry Gelfoam is placed in the epidural
space to protect the dura from muscle attachment during the healing process, which
may lead to postoperative headaches. Reconstruction of the cranial defect is accomplished
using titanium mesh. Advantages of the mesh, aside from producing an excellent cosmetic
result, include its inert properties, ease of implantation, and absence of artifact
on postoperative imaging. There is also no concern of degradation over the years when
using the titanium mesh, as is the case with calcium phosphate cement when there is
incomplete osteogenesis.[27]
Although we did not evaluate each component of our closure technique independently,
many of them have been validated by the neurosurgical literature.[1]
[9]
[15]
[16]
[21]
[26] We believe that it is important to understand each of these closure techniques,
but a primary dural closure should be the goal in all cases. It is a simple and effective
technique and serves as the first step in mitigating CSF leaks after retrosigmoid
craniectomies.
In addition to its benefits in reducing CSF leaks, our closure is efficient and economical.
There are no time-consuming additional steps, such as the harvesting of autograft
(i.e., fat, fascia, galea, etc.) or placement of a lumbar drain. Each of these additional
time-consuming steps also has an associated cost. In 2005, it was estimated that 1
minute of operating room time could cost as much as $133, depending on the procedure
type, and is likely a low estimate today.[28] Costs are also generated with each cranioplasty technique and with the use of dural
sealants and dural allografts; however, it is important to remember that each technique
is utilized to prevent readmissions for CSF leak, which cost nearly $25,000. We believe
our technique minimizes the economic burden while providing maximal benefit to each
patient.
Conclusions
Primary dural closure is possible in retrosigmoid approaches without the use of dural
allo- or autografts. Careful attention to the handling of the dural flap is necessary
to achieve this. This may help obviate the need for graft placement, complete cranioplasty,
or postoperative lumbar drain placement when combined with other well-known closure
techniques.
