The Laryngoscope
Lippincott Williams & Wilkins
© 2007 The American Laryngological,
Rhinological and Otological Society, Inc.
Algorithm for Reconstruction After Endoscopic
Pituitary and Skull Base Surgery
Abtin Tabaee, MD; Vijay K. Anand, MD; Seth M. Brown, MD-MBA; Jerry W. Lin, MD-PhD;
Theodore H. Schwartz, MD
Introduction: The expanding role of endoscopic
skull base surgery necessitates a thorough understanding
of the indications, techniques, and limitations of the var-
ious approaches to reconstruction. The technique and out-
comes of endoscopic skull base reconstruction remain in-
completely described in the literature.
Study Design and Methods: Patients undergoing
endoscopic skull base surgery underwent an algorithmic
approach to reconstruction based on tumor location, de-
fect size, and presence of intraoperative cerebrospinal
fluid (CSF) leak. A prospective database was reviewed to
determine the overall efficacy of reconstruction and to
identify risk factors for postoperative CSF leak.
Results: The diagnosis in the 127 patients in this
series included pituitary tumor in 70 (55%) patients, en-
cephalocele in 16 (12.6%) patients, meningioma in 11
(8.7%) patients, craniopharyngioma in 9 (7.1%) patients,
and chordoma in 6 (4.7%) patients. Successful reconstruc-
tion was initially achieved in 91.3% of patients. Eleven
(8.7%) patients experienced postoperative CSF leak, 10 of
which resolved with lumbar drainage alone. One (0.8%)
patient required revision surgery. Correlation between
postoperative CSF leak and study variables revealed a
statistically significant longer duration of surgery (243 vs.
178 min, P ⫽ .008) and hospitalization (12.1 vs. 4.5 days,
P ⬍ .0001) and a trend toward larger tumors (mean, 3.2
vs. 2.3 cm; P ⫽ .058) in patients experiencing postopera-
tive CSF leak.
Conclusion: The algorithm for reconstruction after
endoscopic surgery presented in this study is associated
with excellent overall efficacy. A greater understanding of
risk factors for postoperative CSF leak is imperative to
achieve optimal results.
From the Department of Otolaryngology–Head and Neck Surgery
(A.T.) Beth Israel Medical Center, Albert Einstein College of Medicine, New
York, New York, U.S.A.; and the Departments of Otorhinolaryngology–
Head and Neck Surgery (V.K.A., S.B., J.W.L.) and Neurological Surgery
(T.H.S.), Weill Medical College of Cornell University, New York, New York,
U.S.A.
Editor’s Note: This Manuscript was accepted for publication March
13, 2007.
Presented at the Triological Society Meeting Combined Sections
Meeting, Marco Island, Florida, U.S.A., February 18, 2007.
Send correspondence to Dr. Vijay Anand, 772 Park Avenue, New
York, NY 10021. E-mail: vijayanandmd.aol.com
DOI: 10.1097/MLG.0b013e31805c08c5
Laryngoscope 117: July 2007
Key Words: Cerebrospinal fluid leak, reconstruction,
endoscopic, skull base, surgery, pituitary.
Laryngoscope, 117:1133–1137, 2007
INTRODUCTION
Endonasal, endoscopic surgery represents an emerg-
ing treatment modality for a variety of pathologies of the
anterior skull base.1 As its popularity expands, a descrip-
tion of the technique and short-term outcomes of the var-
ious aspects of the procedure are required, including the
approach, tumor resection, and reconstruction. The latter
is an integral component of the surgery and is an exten-
sion of the technique described for endoscopic cerebrospi-
nal fluid (CSF) leak closure.2–5 Multiple challenges are
associated with the creation of a watertight separation
between the sinonasal and intracranial cavities including
the heterogeneous nature of the lesions and indirect ac-
cess to the surgical site. However, inadequate closure rep-
resents a major complication with the potential for CSF
leak, meningitis, pneumocephalus, and death. Multiple
reconstruction methods have been described for endo-
scopic pituitary and skull base surgery with various de-
grees of success.6 –12 The current report describes an algo-
rithmic approach for reconstruction based on a large,
prospective case series.
MATERIALS AND METHODS
After institutional review board approval, a prospective da-
tabase of all endoscopic pituitary and skull base procedures per-
formed between January 2004 and December 2006 was reviewed.
All procedures were performed at a single tertiary care medical
center by the senior authors (V.K.A. and T.H.S.). The office, hospi-
tal, and surgical charts were reviewed for patient and tumor
demographics, surgical variables, reconstruction technique, and
incidence of postoperative CSF leak at last follow-up. Exclusion
criteria included incomplete medical records or nonendoscopic or
endoscope-assisted surgery. Potential correlations between study
measures and CSF leak were evaluated using two-tailed Student
t test for continuous variables and ␹2 test and Fisher’s exact test
for variables in proportions. Significance was defined as P ⬍ .05.
Endoscopic Skull Base Reconstruction
All patients underwent reconstruction of the skull base at
the completion of the procedure based on the following algorithm
Tabaee et al.: Reconstruction After Endoscopic Surgery
1133

(Table I). Application of hemostatic agents alone without a formal
reconstruction was performed in patients undergoing extradural
dissection without a bony skull base defect or intraoperative CSF
leak. In patients undergoing closure of an isolated CSF leak/
encephalocele with a small bony defect, a single layer of autolo-
gous fat or fascia was placed as a “bath plug”4 through the defect
and was followed by application of tissue sealant (type I). In
surgery for sellar lesions involving a bony skull base defect with-
out arachnoid violation or intraoperative CSF leak, the cavity
was packed with Gelfoam, and the bony sella was reconstructed
and covered with tissue sealant (type IIA). Reconstruction of the
floor begins with measurement of the defect using a pliable ma-
terial such as a neurosurgical pledget. This is followed by sharp
Laryngoscope 117: July 2007
1134
Fig. 1. Type IIIA multilayered closure in-
corporating packing of tumor cavity with
fat graft (A), underlay placement of fas-
cia (B), onlay placement of bone (C), and
covering of wound edges with tissue
sealant (D).
dissection of the wound edges to promote scar formation and
granulation. The graft (nasal septum vomer or miniplate) is then
trimmed to the desired size and placed as either an underlay or,
more commonly, an onlay graft. If intraoperative CSF leak was
identified in these cases, autologous tissue (abdominal fat) was
used to pack the tumor cavity followed by reconstruction of the
bony sella and application of tissue sealant (Type IIB). Patients
undergoing resection of larger lesions of the sella and anterior
skull base (suprasellar, fovea ethmoidalis, cribriform plate) with
a high-volume intraoperative CSF leak underwent multilayered
closure with an autologous fat graft in the tumor cavity followed
by underlay placement of a fascial layer (fascia latta), onlay
placement of a bony buttress, and application of tissue sealant
Fig. 2. Cross-sectional representation of
type IIIA (A) and type IIIB (B) multilayered
closure for anterior cranial fossa recon-
struction. Fat graft is avoided in patients
with direct communication with ventric-
ular spaces to minimize risk of iatrogenic
hydrocephalus.
Tabaee et al.: Reconstruction After Endoscopic Surgery
 TABLE I.
Results of Endoscopic Skull Base Reconstruction.
Indication
No bony skull base defect
Small CSF leak/bony skull base defect
Sella, no intraoperative CSF leak
Sella with intraoperative CSF leak
Suprasella and anterior cranial fossa
Suprasella and anterior cranial fossa with ventricular communication
CSF ⫽ cerebrospinal fluid.
(type IIIA) (Fig. 1 and 2). In cases involving a direct communica-
tion between the ventricular spaces and the tumor cavity, this
technique was modified such that a fat layer was not packed to
minimize the risk of iatrogenic hydrocephalus (Type IIIB) (Fig. 2).
After reconstruction, Floseal (Baxter Inc., Vienna, Austria)
was routinely placed into the sinonasal cavity. Small folds of
Telfa were placed into the anterior nasal cavity and removed on
the first postoperative day. The sinus cavities were not obliter-
ated, and no formal packing was used. Lumbar drain was used
selectively in patients with a history of active preoperative CSF
leak, if there was concern for increased intracranial pressure
(obesity, hydrocephalus on preoperative imaging), or there was
an anticipated large bony skull base defect. Patients were fol-
lowed during the hospitalization and postoperatively for occur-
rence of CSF leak. In addition to routine questioning for symp-
toms suggestive of CSF leak, patients underwent routine
endoscopic examination and debridement of the surgical site at 1
week, 6 weeks, and 3 months after discharge until the cavity was
well healed. The mean duration of follow-up for this series was
213 days (range, 7 days–2.9 yr; SD, 141 days).
RESULTS
One hundred twenty-seven patients, 77 female (61%)
and 50 male (39%), underwent endoscopic surgery during
the study period. The mean age at the time of surgery was
50 (range, 13– 83; SD, 16) years. The final diagnosis
included secreting pituitary adenoma in 37 (29.1%) pa-
tients, nonsecreting pituitary adenoma in 33 (26.0%)
patients, encephalocele/CSF leak in 16 (12.6%) patients,
meningioma in 11 (8.7) patients, craniopharyngioma in
9 (7.1%) patients, chordoma in 6 (4.7%) patients, Rath-
ke’s cleft cyst in 3 (2.4%) patients, and other in 12 (9.4%)
patients. The mean maximal tumor dimension on pre-
operative imaging was 2.3 (range, 0.4 – 8.0; SD, 1.2) cm.
The surgical approach was based on the location and ex-
tent of the lesion and involved the sella in 71 (55.9%)
patients, tuberculum and planum sella in 22 (17.3%) pa-
tients, cribriform plate and fovea ethmoidalis in 20
(15.7%) patients, clivus in 9 (7.1%) patients, pterygoid in 4
(3.1%) patients, and sphenoid sinus in 1 (0.8%) patient.
The mean duration of the procedure was 184 (range, 73–
405; SD, 78) minutes, and the mean duration of hospital-
ization was 5 (range, 0 –35; SD, 5) days. Planned lumbar
drainage was used in 28 (22.0%) patients.
The skull base reconstruction methods and incidence
of postoperative CSF leak are described in Table I. Eleven
(8.7%) patients experienced postoperative CSF leak, 10 of
which resolved with conservative management consisting
Laryngoscope 117: July 2007
Number of Patients, Postoperative CSF Leak,
Reconstruction n ⫽ 127 (%) n ⫽ 11 (%)
No reconstruction 10 (7.9) 0 (0)
Type I 5 (3.9) 1 (20)
Type IIA 33 (26.0) 0 (0)
Type IIB 39 (30.7) 5 (12.8)
Type IIIA 34 (26.8) 3 (8.8)
Type IIIB 6 (4.7) 2 (33.3)
of lumbar drainage, bed rest, and head of bed elevation for
3 to 5 days. The one (0.8%) CSF leak that required revision
surgery involved a suprasellar approach for a 3 cm me-
ningioma. At the time of revision, complete displacement
of the bony buttress and reconstructive graft into the
nasal cavity was noted. The patient failed a second endo-
scopic approach and ultimately underwent successful
skull base reconstruction through a craniotomy approach.
One (0.8%) graft-related complication was experienced in
this series and involved an abdominal seroma at the fat graft
harvest site. This resolved after aspiration and antibiotic
therapy. There were no incidences of meningitis, intracra-
nial abscess, or other infectious-related complications.
Univariate analysis was performed to explore possi-
ble correlation between study variables and postoperative
CSF leak. A comparison of patients experiencing postop-
erative leak compared with those who did not, respec-
tively, identified a statistically significant longer duration
of procedure (243 vs. 178 min, P ⫽ .008) and hospitaliza-
tion (12.1 vs. 4.5 days, P ⬍ .0001). A trend toward larger
tumor size based on the maximal tumor dimension on
preoperative imaging was noted in patients with postop-
erative CSF leak (mean, 3.2 vs. 2.3 cm; P ⫽ .058). There
were no statistically significant differences between pa-
tients experiencing CSF leak versus those who did not,
respectively, with regard to age (mean, 51 vs. 50 yr), sex
(64% vs. 60% female), body mass index (mean, 27.7 vs.
29.2 kg/m2), surgery extending beyond the sella (54% vs.
43%), location of lesion (42% vs. 55% pituitary origin), or
planned use of lumbar drainage (18% vs. 22%). Although
differences in the rate of postoperative CSF leak were
noted among the different closure types (Table I), limited
numbers precluded meaningful statistical analysis. Of
note, there were no incidences of postoperative CSF leak
in patients who did not have identified intraoperative CSF
leak.
DISCUSSION
Postoperative CSF leak after endoscopic pituitary
and skull base surgery may be associated with significant
morbidity including meningitis, pneumocephalus, and
death.13 The creation of a watertight layer is an integral
component of endoscopic procedures and is potentially
challenging given the variation in size and location of the
defect, indirect accessibility to the surgical site, and
stresses placed on the graft from gravity and intracranial
Tabaee et al.: Reconstruction After Endoscopic Surgery
1135
pressure. As the popularity and indications of endoscopic
pituitary and skull base surgery expand, a description of
the efficacy of reconstruction techniques is required.
The reconstructive philosophy described in this report
has evolved over the past several years and is tailored to the
individual patient. The sequential placement of multiple free
tissue layers, as described in Methods, represents our work-
horse for reconstruction. A stratified approach is used based
on surgical and patient factors including the size, location,
and nature of the lesion, the presence of intraoperative CSF
leak, and patient comorbidities for postoperative CSF leak
(i.e., hydrocephalus). The low rate of postoperative CSF leak
and reconstruction-related complications described in this
series support the safety and utility of this technique.
Identification of patients with a higher inherent risk
of postoperative CSF leak is necessary to identify those
patients who require a robust, multilayered closure and
possibly planned lumbar drainage. Our results suggest
that patients undergoing a potentially complex procedure
(longer surgery time) for a lesion larger than 3 cm may be
at higher risk for postoperative leak. Additionally, pa-
tients who experienced CSF leak required a statistically
longer duration of hospitalization. This is likely because
of, at least in part, the need for additional care directly
related to the leakage including lumbar drainage and bed
rest. Finally, a lack of identified intraoperative CSF leak
strongly predicts a low risk of postoperative CSF leak.
Intraoperative staining of CSF with fluorescein (Akorn
Inc., Buffalo Groove, IL) is useful in this regard and has
been previously described.14 Our protocol involves the in-
trathecal (0.25 mL of injectable 10% solution mixed with
10 mL of CSF) and intravenous injection (0.25 mL of
injectable 10% solution mixed with 3 mL of normal saline)
of fluorescein after premedication with 50 mg of diphen-
hydramine and 10 mg of dexamethasone. The addition of
a blue light filter is useful in differentiating the stained
CSF from the surrounding blood, mucous, and irrigation
fluid in cases with subtle leaks.
Integral to the multilayered closure technique is the
choice of graft materials. This is dependent on multiple
factors including reconstructive needs and surgeon pref-
erence. The ideal graft materials for skull base reconstruc-
tion are 1) available in adequate quantity, 2) biocompat-
ible with minimal chance of resorption, rejection, or
infection, 3) easily harvested with minimal morbidity, 4)
associated with minimal imaging interference, 5) inexpen-
sive, and 6) free of potential disease transmission. The
heterogeneity of the location and size of the skull base
defect requires flexibility in the reconstruction technique
and materials. The successful use of a variety of graft
materials has been described previously.
Our primary choice of nasal septum vomer as the
autologous material for bony skull base reconstruction is
based on its availability during the initial approach, long-
term durability, and avoidance of a separate donor site.
Additional autologous options of cartilage/bone from dis-
tant sites include concha, rib, iliac crest, and calvarium.
Synthetic options include absorbable and nonabsorbable
miniplates that may be similarly placed as onlay or un-
derlay grafts without being drilled or sewn into place.
Although available without additional harvesting, the po-
Laryngoscope 117: July 2007
1136
tential for infection and scatter on imaging are distinct
disadvantages.
Autologous fat is used as our tissue of choice for
filling the tumor cavity because of its accessibility and
postoperative durability. Available sites include the abdo-
men or lateral thigh if fascia is also needed. The creation
of a watertight layer is typically performed with fascia
latta given its availability in adequate amounts, ease of
simultaneous harvest, and minimal patient morbidity. Al-
ternatives include synthetic dura and sinonasal mucosa
(middle turbinate) for small reconstructions.
Our technique involves routine placement of tissue
sealant around the wound edges after insetting of the
reconstructive grafts. Although the sealant will not com-
pensate for an otherwise inadequate reconstruction, it will
allow for graft immobility until mucosalization has oc-
curred. Although a similar effect may be achieved with a
formal packing of the sinonasal cavity, the issues associ-
ated with packing including patient discomfort, bacterial
colonization, tissue adherence, and patient straining dur-
ing pack removal represent major issues.
The role of planned lumbar drainage after endoscopic
pituitary and skull base surgery remains controversial
and poorly defined. Although the potential for shunting of
CSF during the recovery period may theoretically allow
for lower rates of postoperative CSF leak, multiple issues
preclude its universal use, including complications asso-
ciated with prolonged immobility (deep vein thrombosis)
and the drain itself (retained catheter tip). Prior reports
citing endoscopic CSF leak closure rates of up to 97%15
without lumbar drainage support its selective use in en-
doscopic skull base reconstruction. Our current approach
avoids a drain for most procedures and is considered in
patients with large encephaloceles and high-volume CSF
leak preoperatively, with large anterior skull base lesions
such as meningioma, and in cases of anticipated patient
comorbidities including increased intracranial pressure.
Additionally, patients who have postoperative CSF leak
undergo lumbar drainage, bed rest, and head of bed ele-
vation for 3 to 5 days initially. In the 11 cases of postop-
erative leak in this series, this strategy was successful in
10 patients, with 1 patient requiring revision closure for a
displaced graft.
Other approaches to reconstruction after endo-
scopic skull base surgery have been described in the
literature. Kassam et al.7 reported the technical aspects
of a multilayered reconstruction using collagen matrix
underlay graft, acellular dermis onlay graft, oblitera-
tion of the paranasal sinus with a fat graft, and tempo-
rary balloon stenting. The use of various synthetic ma-
terials including collagen fleece, silicone plate, alumina
ceramic, fibrin sealant, and titanium plates have also
been described. The description of a multitude of tech-
niques and materials likely reflects the challenging na-
ture of reconstruction.6 –12
The early results of this series support the efficacy
of multilayered reconstruction with free tissue grafts
after endoscopic pituitary and skull base surgery. The
low rate of postoperative CSF leak and reconstruction-
related complications support the ongoing refinement of
the procedure. Several limitations are identified in this
Tabaee et al.: Reconstruction After Endoscopic Surgery
series, including its uncontrolled, nonrandomized de-
sign and demographic and surgical heterogeneity. Ad-
ditionally, although patients were routinely screened for
postoperative CSF leak by clinical history and endoscopy,
the possibility of occult leak exists and would require
additional objective testing. Finally, the possibility of de-
layed leak would require a longer minimal follow-up to
determine the long-term outcomes. Future studies incor-
porating larger series with extended follow-up may allow
a description of the impact of various patient and surgical
variables on outcomes.
CONCLUSIONS
Reconstruction is a critical component of endoscopic
pituitary and skull base surgery. The potential for signif-
icant CSF–leak-related morbidity mandates a high rate of
successful closure at the time of the initial procedure. The
approach described in this report involves stratification ac-
cording to the needs of the patient and is associated with an
acceptably low rate of postoperative CSF leak. Certain fac-
tors such as duration of the procedure and size of the lesion
may be associated with higher risk of CSF leak. As the
indications and popularity of endoscopic pituitary and skull
base surgery expand, large case series with outcomes data
are critical to support future development.
BIBLIOGRAPHY
1. Mehta RP, Cueva RA, Brown JD, et al. What’s new in skull
base medicine and surgery? Skull Base Committee Report.
Otolaryngol Head Neck Surg 2006;135:620 – 630.
2. Zweig JL, Carrau RL, Celin SE, et al. Endoscopic repair of
cerebrospinal fluid leaks to the sinonasal tract: predictors
of success. Otolaryngol Head Neck Surg 2000;123:195–201.
3. McMains KC, Gross CW, Kountakis SE. Endoscopic manage-
ment of cerebrospinal fluid rhinorrhea. Laryngoscope 2004;
114:1833–1837.
4. Wormald PJ, McDonogh M. The bath-plug closure of anterior
skull base cerebrospinal fluid leaks. Am J Rhinol 2003;17:
299 –305.
Laryngoscope 117: July 2007
5. Tabaee A, Kassenoff TL, Kacker A, Anand VK. The efficacy of
computer assisted surgery in the endoscopic management
of cerebrospinal fluid rhinorrhea. Otolaryngol Head Neck
Surg 2005;133:936 –943.
6. Cappabianca P, Cavallo LM, Esposito F, Valent V, de Divitiis
E. Sellar repair in endoscopic endonasal transsphenoidal
surgery: results of 170 cases. Neurosurgery 2002;51:
1365–1372.
7. Kassam A, Carrau RL, Snyderman CH, Gardner P, Mintz A.
Evolution of reconstructive techniques following endo-
scopic expanded endonasal approaches. Neurosurg Focus
2005;19:1–7.
8. Hadad G, Bassagasteguy L, Carrau RL, et al. A novel recon-
structive technique after endoscopic expanded endonasal
approaches: vascular pedicle nasoseptal flap. Laryngoscope
2006;116:1882–1886.
9. Leong JL, Citardi MJ, Batra PS. Reconstruction of skull base
defects after minimally invasive endoscopic resection of
anterior skull base neoplasms. Am J Rhinol 2006;20:
476 – 482.
10. Cappabianca P, Cavallo LM, Valente V, et al. Sellar repair
with fibrin sealant and collagen fleece after endoscopic
endonasal transsphenoidal surgery. Surg Neurol 2004;62:
227–233.
11. Seiler RW, Mariani L. Sellar reconstruction with resorbable
Vicryl patches, gelatin foam, and fibrin glue in transsphe-
noidal surgery: a 10-year experience with 376 patients.
J Neurosurg 2000;93:762–765.
12. Cappabianca P, Cavallo LM, Mariniello G, de Divitiis O,
Romero AD, de Divitiis E. Easy sellar reconstruction in
endoscopic endonasal transsphenoidal surgery with
polyester-silicone dural substitute and fibrin glue: techni-
cal note. Neurosurgery 2001;49:473– 475.
13. Black PM, Zervas NT, Candia GL. Incidence and manage-
ment of complications of transsphenoidal operation for pi-
tuitary adenomas. Neurosurgery 1987;20:920 –934.
14. Keerl R, Weber RK, Draf W, Wienke A, Schaefer SD. Use of
sodium fluorescein solution for detection of cerebrospinal
fluid fistulas: an analysis of 420 administrations and re-
ported complications in Europe and the United States.
Laryngoscope 2004;114:266 –272.
15. Casiano RR, Jassir D. Endoscopic cerebrospinal fluid rhinor-
rhea repair: is a lumbar drain necessary? Otolaryngol
Head Neck Surg 1999;121:745–750.
Tabaee et al.: Reconstruction After Endoscopic Surgery
1137