Peer-Review Reports
The Endoscopic Endonasal Approach to Repair of Iatrogenic and Noniatrogenic Cerebrospinal Fluid Leaks and Encephaloceles of the Anterior Cranial Fossa Jeffrey C. Bedrosian2, Vijay K. Anand2, Theodore H. Schwartz1,3
Key words Anterior skull base - Endoscopic - Minimal access - Minimally invasive - Spontaneous CSF leak - Traumatic CSF leak -
Abbreviations and Acronyms BIH: Benign intracranial hypertension BMI: Body mass index CSF: Cerebrospinal fluid CT: Computerized tomography FESS: Functional endoscopic sinus surgery HRCT: High resolution CT ICP: Intracranial pressure MRI: Magnetic resonance imaging VP: Ventriculoperitoneal Departments of 1Neurological Surgery, 2 Otolaryngology—Head and Neck Surgery, 3 Neurology and Neuroscience, Weill Cornell Medical College, New York Presbyterian Hospital, New York, New York, USA To whom correspondence should be addressed: Theodore H. Schwartz, M.D. [E-mail: [email protected]] Citation: World Neurosurg. (2014) 82, 6S:S86-S94. http://dx.doi.org/10.1016/j.wneu.2014.07.018 Journal homepage: www.WORLDNEUROSURGERY.org
- OBJECTIVE:
The current approach for the diagnosis and repair of spontaneous and traumatic anterior skull-base defects is oulined, highlighting the controversies that exist in the field and describing the strategies required to access different segments of the anterior cranial fossa.
- METHODS:
We reviewed the literature concerning endoscopic management of anterior skull-base defects. These publications have been combined with our own experience repairing cerebrospinal fluid (CSF) leaks and encephaloceles that developed spontaneously, traumatically, or intentionally as a result of endoscopic skull-base surgery.
- RESULTS:
We present a systematic methodology for the repair of these defects. We have divided our surgical approach into four separate corridors. These are the transnasal, transsphenoidal, transethmoidal, and transmaxillary corridors. Dissection strategies vary for each corridor, but with a combination of approaches, all areas of the anterior skull base can be accessed. Skull-base defects are successfully repaired with a multilayered closure that often involves use of a vascularized pedicled mucosal flap. Adoption of this technique has decreased our rate of postoperative CSF leak from 5.9%e3.1%.
- CONCLUSIONS:
Endoscopic endonasal repair of CSF leaks and encephaloceles has evolved significantly during the past decade. The versatility of different endoscopic approaches through the four endonasal corridors allows for the endoscopic repair of almost all skull-base defects. The use of vascularized pedicled mucosal flaps has evolved to cover these defects as part of multilayered closure strategies.
Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2014 Elsevier Inc. All rights reserved.
INTRODUCTION Noniatrogenic cerebrospinal fluid (CSF) leaks occur commonly in the anterior cranial fossa. Iatrogenic CSF leaks can occur inadvertently during endoscopic sinus surgery and as a natural course of endoscopic skull-base surgery. For spontaneous CSF leaks, primary closure rates have improved from 70%e80% in the 1980s (16) to >90% with the evolution of endoscopic techniques (14, 23, 25, 30). As such, endoscopic endonasal repair of CSF leaks and encephaloceles has begun to replace traditional open craniotomy techniques. Closure rates for iatrogenic high-flow CSF leaks that occur during endoscopic skull-base surgery have improved from 20%e70% in early publications to 25 cm H2O and normal biochemical and cytologic composition of CSF 6. No other explanation for the increased intracranial pressure CT, computerized tomography; MRI, magnetic resonance imaging; LP, lumbar puncture; CSF, cerebrospinal fluid.
lateral defects may occur due to improper fusion of Sternberg’s canal, a phenomenon discussed later. In addition, these encephaloceles tend to be quite large, as increased ICP pushes intracranial contents
ENDOSCOPIC CSF AND ENCEPHALOCELE REPAIR
through relatively small skull-base defects. Spontaneous CSF leaks are also highly associated with the female gender and obesity. In one series of 55 consecutive patients with spontaneous CSF leaks, 70% of patients were women and 46 of 55 patients were obese, with an average body mass index (BMI) of 36.2 kg/m2. Elevated ICP persisted in these patients postoperatively, with an average lumbar drain pressure of 27 cm H2O (39). In another series of 21 patients, 18 patients were women, with an average BMI of 31.2 kg/m2 (40). In many ways, spontaneous CSF leaks are the most difficult to treat. Small, slow leaks can be hard to locate. Broad attenuation of the bone of the skull base makes achieving a watertight closure difficult. Associated encephaloceles or meningoencephaloceles are found in 50%e100% of patients. Postoperative recurrence rates are generally higher compared with all other causes, ranging from 25%e87% in some series (16, 27, 30). Important, understanding the etiology and demographics of spontaneous CSF leaks is critical to
successful primary closure. Although repair techniques are similar to the closure of other leak types, postoperative ICP management of these patients is critical. The details and controversies of postoperative ICP management will be discussed later. Traumatic CSF leaks may be divided into iatrogenic and noniatrogenic leaks. Iatrogenic CSF leaks most commonly occur during transsphenoidal pituitary tumor resection (0.5%e15% incidence). They also frequently occur during acoustic neuroma surgery (7%e11% incidence) and during functional endoscopic sinus surgery (0.5%e3% incidence) (21). The most common sites of inadvertent skull base violation during endoscopic sinus surgery are the lateral lamella of the cribriform plate and the roof of the posterior ethmoid sinuses (3), as the skull base slopes inferiorly toward the face of the sphenoid sinuses (Figure 2). Of these leaks, 50% present intraoperatively or immediately postoperatively, whereas the remaining 50% are delayed, occurring 1 week to 1 month postoperatively. Theories for delayed iatrogenic CSF leak presentation include wound contraction, flap devascularization, or necrosis, resolving cerebral edema and increased ICP (21). Traumatic noniatrogenic CSF leaks are usually due to accidental trauma (70%e 80% incidence). Two to 4% of all acute head injuries result in CSF rhinorrhea. In 70% of cases, these leaks resolve spontaneously with observation or a lumbar drainage; however, repair should be considered if leaks do not resolve within 1e2 weeks to prevent meningitis (21). Intracranial tumors may cause CSF leaks as well. Tumor growth may obstruct CSF reabsorption leading to increased ICP and a high pressure leak. Alternatively, intracranial or extracranial tumor growth may directly erode the skull base. LOCALIZING THE DEFECT
Figure 1. Artist’s illustration of the fovea ethmoidalis (inset), a frequent site for inadvertent skull base violation. A sagittal view of the skull base demonstrating possible skull base defect locations and encephalocele formation. Defects shown are at the foramen cecum (anterior) and the posterior cribriform region (posterior) as the skull base slopes downward to meet the face of the sphenoid sinus.
WORLD NEUROSURGERY 82 [6S]: S86-S94, DECEMBER 2014
Preoperative and intraoperative localization of a CSF leak is an essential component of successful repair. There are several options for preoperative localization. High resolution CT (HRCT), CT cisternography, magnetic resonance imaging (MRI), MRI cisternography, and radionuclide cisternography will be discussed. HRCT is ideally noninvasive; it does not require any
www.WORLDNEUROSURGERY.org
S87
PEER-REVIEW REPORTS JEFFREY C. BEDROSIAN ET AL.
ENDOSCOPIC CSF AND ENCEPHALOCELE REPAIR
Figure 2. An iatrogenic defect of the fovea ethmoidalis (crosshairs) is shown. A small encephalocele is present.
intrathecal injection. HRCT is capable of resolving millimeter-sized skull-base defects and as such, its chief drawback is the detection of congenital skull-base defects not associated with CSF leak or encephalocele. HRCT also exposes patients to relatively high doses of radiation. In 1 study, high resolution CT was used in 40 patients with anterior skull-base defects. Forty-two defects were detected with HRCT. The investigators correlated their findings with intraoperative image guidance, achieving a primary endoscopic endonasal repair rate of 95% (40/42 defects). This rate improved to 100% after revision surgery (41). CT cisternography requires injection of contrast material intrathecally with a LP. This technique is 85% sensitive for detecting active leaks at the time of the study, but its sensitivity varies between 48% and 96% for inactive leaks. Contrast material does not distribute evenly throughout the CSF and the small quantity
S88
www.SCIENCEDIRECT.com
of contrast passing through a hairline leak may not be able to be resolved from the surrounding bone (33). CT cisternography is useful when the frontal sinus or sphenoid sinus serves as a reservoir for accumulating CSF fluid, as these reservoirs will demonstrate contrast enhancement (21). The drawbacks of CT cisternography are its radiation exposure, especially if delayed scans are required, its invasive nature, possible contrast allergic reactions, and rarely intracranial hemorrhage. T2-weighted MRI has been used to detect high signal intensity fluid communicating between the sub-arachnoid space and the paranasal sinuses or encephalocele herniation into the sinuses. This technique lacks specificity, however, and may be confused with preexisting paranasal sinus inflammation, leading to an unacceptably high false-positive rate. MRI cisternography requires gadolinium injection into the subarachnoid space. In a prospective study of 20 patients with both
active and intermittent CSF rhinorrhea, a CSF leak was localized in 16 patients after MRI cisternography with gadolinium. T1 fat-suppressed images were taken before and after injection in a 30- to 40-degree prone head-down position during a 15- to 30-minute period. Aiden et al. (1) reported no adverse side effects from the 0.5-mL intrathecal gadolinium injection. This technique, although not yet widely adopted, offers a promising alternative to HRCT or CT cisternography and their relatively high levels of radiation exposure. MRI imaging is also useful for the differentiation of meningoceles and meningoencephaloceles. In another study using MRI cisternography, Algin et al. (2) reported a 100% sensitivity for detection of CSF leakage. In a larger study of 85 patients, MRI cisternography detected a CSF leak in 100% of patients with clinically evident meningitis or continuous CSF rhinorrhea (33). CSF leak was detected in 70% of patients with confirmed intermittent leakage and 60% of patients with suspected CSF rhinorrhea (33). Radionuclide cisternography may be conducted during several days and is ideal for low volume intermittent leaks. It uses technetium 99 administered intrathecally. Its chief drawback is its poor spatial resolution compared with CT and MRI cisternography. Intranasal pledgets may be placed endoscopically to help localize the leak. Radionuclide cisternography has a high 33% false-positive rate, with a sensitivity that varies between 62% and 76%. For intraoperative localization, one can use intrathecal fluorescein. The safety of routine use of intrathecal fluorescein, an off-label application of the drug, has been debated in the literature. Our protocol for administering intrathecal fluorescein is as follows: informed consent for intrathecal fluorescein injection is obtained. After intubation, 10 mg of dexamethasone and 50 mg of diphenhydramine are administered to the patient. A LP is performed, with or without placement of a lumbar drain. A total of 10 mL of CSF is withdrawn and mixed with 0.25 mL of 10% fluorescein. The CSF/fluorescein mixture is slowly reinjected into the intrathecal space. In our own retrospective review of 54 patients undergoing endoscopic skull-base surgery, 3 of 54 patients experienced postoperative lower extremity weakness or numbness. These results were possibly confounded by
WORLD NEUROSURGERY, http://dx.doi.org/10.1016/j.wneu.2014.07.018
PEER-REVIEW REPORTS JEFFREY C. BEDROSIAN ET AL.
placement of a lumbar drain in 2 of the 3 patients (28). In a larger series (18) of 203 pituitary adenomas, there were no clear fluorescein-related complications. Intrathecal fluorescein is an invaluable tool for detection of intraoperative CSF leak and is highly predictive of postoperative CSF leak. In a retrospective analysis (38) of 61 patients undergoing endoscopic skull-base surgery for a variety of lesions, 7 of 37 patients (19.4%) with an intraoperative fluorescein leak experienced postoperative CSF leak. Of the remaining 24 patients, no fluorescein leak was visualized. There was a 0% incidence of postoperative leak among these patients (38). With respect to repair of preexisting skull-base defects, identification of intraoperative fluorescein leak is essential. For small leaks, use of a blue light filter attached to the endoscope greatly enhances the sensitivity of detection of fluorescein. With this technique, fluorescein may be visually resolved at 1 part in 10 million (21). In a retrospective case series of 103 patients undergoing CSF leak repair, intrathecal fluorescein was used in 47 cases. Most were cases of spontaneous CSF leak (61.7%). Of these patients, fluorescein was visualized at the leak site in 66% of cases. A leak was identified without visualization of fluorescein in 23.4% of cases (false negatives). In 10.6% of cases (5 patients), no leak was identified and no fluorescein was identified. Nonetheless, the most likely site of leak was repaired. Of these 5 patients, 2 developed postoperative recurrent CSF leak. Fluorescein sensitivity for detecting an intraoperative CSF leak was 73.8% and specificity was 100%. The rate of recurrent leak was 31.3% when fluorescein was not visualized compared with 9.7% when fluorescein was visualized; however, this finding was not statistically significant (P ¼ 0.10) (35). ADVANTAGES AND DISADVANTAGES OF THE ENDOSCOPIC ENDONASAL APPROACH Endoscopic, endonasal skull-base approaches have significant advantages when compared with traditional transcranial open surgery. Endoscopic approaches do not require brain retraction or neurovascular manipulation. They do not require craniotomies and large scalp incisions. They do not result in a cosmetic deformity. As a result, the complication
ENDOSCOPIC CSF AND ENCEPHALOCELE REPAIR
rate and hospital stay after these surgeries is significantly decreased. In addition, when compared with the microscope, the endoscope offers a wider, more panoramic view of the surgical field. This is due to the position of the endoscope lens in close proximity to the operative site, versus a microscope lens several centimeters more distant. By the same principle, the endoscope offers superior lighting of the operative field compared with the microscope. Endoscopes can also have angled lenses on their tips, providing the ability to look around corners and operate in areas obscured from the line-of-site limitations of the microscopic view. Advances in video screen and camera technology allow the endoscopic surgeon to enjoy a high-definition view of the surgical field that is comparable to the clarity observed through the oculars of a microscope. The principle drawback of endoscopic surgery is the lack of binocular vision provided by the microscope. Otolaryngologists tend to be familiar and comfortable with this 2-dimensional endoscopic picture because of their extensive experience with endoscopic sinus surgery. Indeed, 3 dimensionality can be approximated by continued movement of the endoscope in and out of the nose, as surgical landmarks change size on the screen relative to their proximity to the endoscope lens. For neurosurgeons used to working through the microscope, performing microsurgery in a 2-dimensional world requires acclimation and is associated with a learning curve. Within the past few years, however, 3dimensional endoscopes have become commercially available. Although not widely used, these new endoscopes are capable of displaying the 3-dimensional contours of the skull base, sinonasal cavity, and sella (6, 13, 37). At present, these endoscopes do not offer the same highdefinition resolution as the standard endoscopic view, and they require the surgeon to wear 3-dimensional glasses. Nonetheless, as the technology improves, use of 3-dimensional endoscopes will undoubtedly become more widespread, with visual clarity and depth of field approaching the microscope’s standard. ENDONASAL OPERATIVE APPROACHES CSF leaks can occur anywhere throughout the anterior cranial fossa. As such, a
WORLD NEUROSURGERY 82 [6S]: S86-S94, DECEMBER 2014
variety of endoscopic approaches are necessary to access these disparate regions of the skull base. In general, the degree of dissection and degree of difficulty required increases as one moves further away from the midline; however, the basic principles of successful endoscopic closure always apply. We have previously classified our broad array of endoscopic approaches for resection of skull-base tumors based on 1) an anatomic target, 2) a cranial base approach, and 3) a nasal corridor. This scheme can also be applied to strategies for repair of CSF leaks and encephaloceles (32). We identified 12 separate anatomic skullbase targets. Of those, the ones likely associated with a CSF leak include the anterior fossa, olfactory groove, sella, suprasellar cistern, and pterygopalatine fossa. These regions can be approached through four possible corridors: transnasal, transsphenoidal, transethmoidal, and transmaxillary (Figure 3). Understanding these corridors anatomically is necessary for appropriate and safe dissection performed by the otolaryngologist during the approach to the skull-base defect target. In this context, we define the endonasal approach as a surgical strategy based on the nasal corridor traversed on our way to the target. Therefore, it is the specific dissection technique that gets us from the nostril to the target. Each approach applicable to the repair of a potential CSF leak will be described in this manner. The transcribriform approach is the most midline approach. Access to the cribriform plate is gained through the transnasal corridor. The transnasal corridor is defined as those spaces within the nose and skull base accessed without transgression of the sinuses. Strictly speaking, all endonasal procedures begin in the transnasal corridor, but midline structures like the olfactory groove and cribriform plate may be accessed without further dissection. The lateral boundaries of this approach are the middle turbinates bilaterally. Encephaloceles and meningoencephaloceles often herniate into this region of the nose. Often it is necessary to perform a partial septectomy, removing the perpendicular plate of the ethmoid bone to circumscribe these lesions by operating through both nostrils. Doing so, however, will invariably damage the olfactory mucosa, which extends inferiorly from
www.WORLDNEUROSURGERY.org
S89
PEER-REVIEW REPORTS JEFFREY C. BEDROSIAN ET AL.
ENDOSCOPIC CSF AND ENCEPHALOCELE REPAIR
Figure 3. An artist’s illustration of the four sinonasal corridors with the skull in three-quarter projection.
the cribriform plate onto the superior turbinates. This leads to anosmia. This basic approach exposes a narrow segment of the skull base centered on the crista galli. When more lateral access is required, a transfovea ethmoidalis approach is required. The fovea ethmoidalis is a common site of accidental skull base transgression during endoscopic sinus surgery (Figure 4). The foveal ethmoidalis is accessed through the transethmoidal corridor. In this corridor, the middle turbinates are now the medial boundaries
Figure 4. The fovea ethmoidalis is a frequently violated anatomic landmark during endoscopic sinus surgery (arrow).
S90
www.SCIENCEDIRECT.com
and the lamina paprecea of the orbits are the lateral boundaries of the dissection. Otolaryngologists are used to developing the transethmoidal corridor during a complete ethmoidectomy, a routine element of functional endoscopic sinus surgery (FESS). This requires removal of the uncinate process, a thin vertically oriented bone that visually obscures the natural ostium of the maxillary sinus and the ethmoid bulla. The ethmoid bulla is the anterior-most ethmoid air cell and its removal is the first step in a complete ethmoidectomy. Identification of the skull base so that it can be avoided is of primary importance during this operation. Nonetheless, the lateral lamella of the fovea is the thinnest bone of the skull base and is easily violated. Such iatrogenic defects are usually noticed at the time of sinus surgery. If such an injury occurs, immediate repair is advised. In this case, the surgical corridor is already present as part of the dissection for the FESS. Development of this corridor in combination with the transnasal corridor provides lamina to the lamina skull base access in the coronal plane. It also provides access from the
bony beak of the frontal recess to the face of the sphenoid sinus in the sagittal plane. CSF leaks may be in the sphenoid sinus. They can occur as a result of iatrogenic accidental trauma, or intentionally, after removal of sellar, suprasellar, or cavernous sinus lesions. They can occur spontaneously as a result of incompetence of the diaphragma sellae with meningoencephalocele herniation leading to primary empty sella, or even empty sella syndrome. Alternatively, far lateral sphenoid sinus leaks may occur due to congenital skull base dehiscence in this region. In all patients, the transsphenoidal corridor provides access to the sphenoid sinus. The anatomic targets reached through the transsphenoidal corridor are the sella and the suprasellar sistern, accessed with the transsellar and transtuberculum approach, respectively. The cavernous sinus and planum sphenoidale may be accessed with these approaches as well. The endonasal corridor provides access to the face of the sphenoid sinus. The natural ostia of the sphenoid sinuses can be located by directing the endoscope medially to the middle turbinate. These ostia are then widely enlarged bilaterally. If one wishes to work in the sphenoid sinuses bimanually, a posterior septectomy can be performed with or without preservation of a nasoseptal flap (discussed later). The intersinus septum can then be drilled away allowing access to the breadth of the sphenoid sinus by both the endoscope and surgical instruments necessary to repair the leak. If the leak is located in the suprasellar region, the planum sphenoidale and tuberculum sella may be drilled away once the transsellar corridor has been created. To achieve adequate visualization and instrumentation of all these transsphenoidal approaches, it is essential to widely open the face of the sphenoid sinus bilaterally. This may require removal of both superior turbinates. The intersinus septum and any partial sphenoid septations should be drilled flush with the face of the sella and opticocarotid recess with a diamondtipped burr. Only after the completion of these maneuvers can the entire height and breadth of the sphenoid sinus be appreciated (32). Lateral sphenoid sinus CSF leaks may occur as a result of a dehiscence of Sternberg’s canal. During development,
WORLD NEUROSURGERY, http://dx.doi.org/10.1016/j.wneu.2014.07.018
PEER-REVIEW REPORTS JEFFREY C. BEDROSIAN ET AL.
each half of the sphenoid bone requires the fusion of five cartilaginous precursors. Normal development of the greater wing of the sphenoid requires complete fusion with the presphenoid and basisphenoid precursors. When this does not occur, a lateral craniopharyngeal canal (Sternberg’s canal) remains (Figure 5). This skull base dehiscence occurs in the lateral recess of the sphenoid sinus and is not adequately visualized through the transsphenoidal corridor (36). A CSF leak or meningoencephalocele has been reported to occur through this dehiscence (Figure 6). At present, several case series (10, 29, 36) have been published addressing the treatment of this rare clinical entity. Endoscopic access to the far lateral sphenoid requires a combination of the transsphenoidal, transethmoidal, and transpterygoid approaches. The transpterygoid approach is accessed by the transmaxillary corridor. The transmaxillary corridor is developed by identification of the natural ostium of the maxillary sinus after removal of the uncinate process. The maxillary ostium can then be widely enlarged using standard FESS techniques to visualize the posterior wall of the maxillary sinus. The posterior wall of the maxillary sinus may then be drilled away, providing access to the pterygopalatine fossa and its neurovascular contents. The sphenopalatine artery is then identified and ligated. The remaining neurovascular structures of the pterygopalatine fossa can then be reflected laterally so that the pterygoid process may be drilled down. This maneuver provides access to the lateral recess of the sphenoid sinus, revealing the origin of Sternberg’s canal. When this approach is combined with the transethmoidal, transsphenoidal approaches, the sphenoid sinus can be widely visualized and instrumented. Using
Figure 5. An artist’s illustration of the sphenoid bone. The arrow points to a lateral dehiscence of Sternberg’s canal.
ENDOSCOPIC CSF AND ENCEPHALOCELE REPAIR
this technique, Tabaee et al. (36) presented a case series of 13 patients with far lateral meningoencephaloceles. In their series 11 of 13 patients were successfully closed after 1 surgical attempt. Two patients developed postoperative CSF rhinorrhea. Of these, 1 postoperative leak resolved spontaneously and the other required reoperation. All patients were followed for an average of 4.7 years ( 3.3 years) without leak recurrence. Facial numbness is a potential morbidity of this approach, given the proximity of the infraorbital nerve within the pterygopalatine foramen. In their series, Tabaee et al. (36) reported 1 of 13 patients with postoperative facial paresthesia. There were no reported complications in a series of 15 patients by Castelnuovo et al. (10). THE ENDOSCOPIC ENDONASAL APPROACH TO THE FRONTAL SINUSES The frontal sinuses are the most difficult area of the skull base to reach endoscopically. CSF leaks and encephaloceles of the frontal sinuses have traditionally been repaired through a pure open approach with the creation of an osteoplastic flap, or through combined endoscopic and open approaches. It may be difficult to visualize and repair CSF leaks of the far lateral frontal sinus through endoscopic means alone. Even if a defect can be visualized, a complex multilayered repair is usually not possible. In addition, the narrow frontal sinus outflow tract makes placement of a vascularized pedicled mucosal flap impossible. In their series of 28 consecutive patients with 32 various skull-base leaks, Nyquist et al. (25) reported only 2 failures after endoscopic repair. One of these failures, however, was located in the posterior wall of the frontal sinuses. This recurrence, which occurred after a vehicular trauma, was identified adjacent to a previously repaired large meningoencephalocele defect site. A transcranial approach for the revision surgery was chosen due to the risk of frontal sinus outflow tract stenosis. Becker et al. (5) reported 2 cases of purely endoscopic endonasal repair of frontal sinus CSF leaks. One patient developed a CSF leak of the lateral frontal sinus. This lesion was visualized with a frontal sinus drillout (Draf III/Modified Lothrop procedure) and creation of a
WORLD NEUROSURGERY 82 [6S]: S86-S94, DECEMBER 2014
Figure 6. The lateral dehiscense of Sternberg’s canal is usually not visualize through the sphenoid sinus. Instead a transpterygoid fossa approach is required to identify the encephalocele.
25 35 mm frontal sinusotomy. A fat graft was placed to seal the defect; however, the leak recurred 19 days later. This recurrence was successfully treated with lumbar drain placement for 7 days. This group also reported successful primary repair of a 5 7 mm superior frontal sinus defect using a similar technique (5). ENDOSCOPIC ENDONASAL CLOSURE TECHNIQUES Spontaneous CSF Leaks Various graft materials and multilayer closure strategies have been successfully used to seal skull-base defects; however, there are several basic principles that are universally applicable. First, the defect must be circumferentially visualized. If an encephalocele or meningoencephalocele is present, it should be reduced or cauterized flush with the skull base, as herniated brain tissue is typically nonfunctional. Next, the skull base should be circumferentially demucosalized around the defect site. There should be no overlap between the in situ skull-base mucosa and the mucosal graft applied during the closure. Any such overlap may result in development of a mucocele that will expand with time and require a further operation to correct. Finally, it is critical that the mucosal overlay graft lie flush with the skull base. Occasionally, the defect may need to be recontoured so that it is flat enough to allow the graft to make adequate contact with the skull base (4, 31).
www.WORLDNEUROSURGERY.org
S91
PEER-REVIEW REPORTS JEFFREY C. BEDROSIAN ET AL.
Grafting may be performed with an underlay technique, with an overlay technique, or with a combination of both techniques. Underlay grafting require placement of a graft between the dura and the skull base, resting against the cranial side of the skull base. Options for underlay graft materials include septal mucoperiosteum, temporalis fascia, Duragen, cadaveric fascia, dermis, or pericardium. Closure rates are not affected by the type of underlay used (31). All grafts should circumferentially overlap all edges of the defect, resting flush with the cranial surface of the skull base as part of a multilayered closure. Placement of an underlay graft for defects smaller than 4 mm may be difficult. In these cases, overlay grafting alone is usually sufficient. We often interpose a fat graft between these small defects and our mucosal graft, sealing the multilayer closure in place with Duraseal. If overlay grafting is needed, it may be performed with either a free mucosal graft or a vascularized pedicled mucosal flap. Free mucosal grafts are well-described in the literature. Sources for these grafts often come from adjacent septal or inferior turbinate mucosa. Vascularized pedicled mucosal flaps may be derived from the nasoseptal mucosa, the middle turbinate mucosa, the inferior turbinate mucosa, or the nasal floor mucosa. Choice of the source of the flap depends on the size and location of the defect to be covered. The nasoseptal flap is the most popular vascularized pedicled flap (15). It is the easiest to harvest, has a robust blood supply, and offers the largest coverage and the longest pedicle. This flap is based on the nasoseptal artery, a terminal branch of the sphenopalatine artery. The pedicle for the flap spans the distance between the natural ostium of the sphenoid sinus and the ipsilateral apex of the choana. The septal mucosa can be harvested quite anteriorly, stopping at a point that approximates the anterior tip of the inferior turbinate (20). For very large skull-base defects, two nasoseptal flaps can be harvested from each side of the nasal septum. This technique has been termed the Janus flap as the two flap faces may be onset adjacent to each other to cover a large anteroposterior defect (26). Iatrogenic CSF Leaks Transsphenoidal excision of pituitary macroadenomas accounts for about half of the
S92
www.SCIENCEDIRECT.com
ENDOSCOPIC CSF AND ENCEPHALOCELE REPAIR
skull-base tumors removed in our practice. In our experience, the rate of iatrogenically created intraoperative CSF leak rises as tumor size increases (18). As previously described, intraoperative CSF leak is easily verified by preoperative injection of intrathecal fluorescein. After tumor removal, if a CSF leak is present, the sellar defect is closed in a multilayered fashion. After hemostasis is achieved, a free fat graft, usually taken from the abdomen, is placed in the defect. Care is taken not to over-pack the fat in the defect to prevent compression of the normal pituitary. Medpore is then sized to be slightly larger than the bony sellar defect and inset into the sella to support the fat graft. An overlay vascularized pedicled flap is then applied and the closure is sealed with Duraseal. Adoption of the vascularized pedicled flap for the multilayered closure of sellar defects decreased our postoperative CSF leak rate from 5.9%e2.9% in our own series of 135 patients (Table 2) (24). For large skull-base defects that are intentionally created during the removal of nonpituitary skull-base tumors, we have developed a gasket seal closure technique (22). After circumferentially defining the defect, a fascia lata graft is harvested. The dimensions of this graft are at least 1 cm larger on all sides than the size of the defect. This allows the graft to be partially inset into the defect. The graft is held in place with Medpore, sized to rest on the cranial surface of the skull base in an underlay fashion. In this way, the fascia lata creates a watertight seal, as parts of the graft lie both intracranially and extracranially. Intracranial pressure holds this gasket seal tightly against the skull base and a vascularized pedicled mucosal graft completes the closure in an overlay fashion (22). We analyzed our own series of 415 skull base cases to determine the benefit of the nasoseptal flap for the prevention of postoperative CSF leak. Of the 301 cases, 205 were performed prior to adoption of the nasoseptal flap and 210 cases were performed after the adoption of the nasoseptal flap in March 2008. Of these 210 patients, 96 cases had a nasoseptal flap (Table 2). All four endonasal corridors were represented in our series. After March 2008, our rate of postoperative CSF leak decreased by approximately one-half, from 5.9% (before March 2008) to 3.1% (after March 2008 nasoseptal flap group) and 2.6% (after March 2008 non-nasoseptal
flap group). Our improved non-nasoseptal flap closure rate after March 2008 can be explained by more selective use of this technique for smaller skull-base defects, usually without an intraoperative CSF leak. POSTOPERATIVE MANAGEMENT Postoperative antibiotic use to prevent meningitis is controversial. We use antibiotics with Gram-positive coverage for as long as nasal packing remains. However, nasal packing is typically removed within the first 24 hours postoperatively. Telfa nasal packing placed in both nares is used to minimize postoperative nasal bleeding and also to bolster our closure while the Duraseal placed over the mucosal flap hardens. For nonpituitary skull-base tumors, triple antibiotic coverage with vancomycin, metronidazole, and a second or third generation cephalosporin is used for the first 72 hours. Despite the nonsterile nature of the nose, postoperative intracranial infection is rare. In a comprehensive review of their experience with 800 skull base patients, Kassam et al. (19) reported a 1.9% rate of postoperative infection confirmed by bacterial culture. Among these, 13 cases resulted in meningitis, 1 case resulted in an intradural abscess, and 1 case resulted in an extradural abscess. For purely endoscopic pituitary resections, the rate of postoperative infection has been reported to be even lower (0.7%e1.0%) in two large series (7, 11). These postoperative infection rates are significantly decreased when compared with traditional open approaches (12, 34) despite the nonsterile nature of the nose. We speculate that the low rate of postoperative infection is due to the relatively small skull base opening of the endoscopic approach when compared with traditional open craniotomies. Postoperative management of elevated ICP in patients with high pressure tumorrelated leaks or spontaneous leaks with BIH can help prevent leak recurrence (39). After surgery, the head of bed is elevated to 30 degrees. Patients are placed on stool softeners and instructed to avoid nose blowing, drinking through a straw, sneezing with their mouth closed, or lifting objects heavier than 5 pounds for the next 2 weeks. Other adjunctive measures for the management of postoperative elevated ICP can include placement of a lumbar drain, placement of a
WORLD NEUROSURGERY, http://dx.doi.org/10.1016/j.wneu.2014.07.018
PEER-REVIEW REPORTS JEFFREY C. BEDROSIAN ET AL.
ENDOSCOPIC CSF AND ENCEPHALOCELE REPAIR
Table 2. Our Series of 415 Skull Base Patients Before and After the Adoption of the Nasoseptal Flap After Adoption of NSF Reconstruction in March 2008
Total number of patients Men Age, mean (SD) (years) Prior surgery
Group A (NSF), (n, %)
Group B (non-NSF), (n, %)
Before Adoption of NSF Reconstruction in March 2008 (group C), (n, %)
96 (100)
114 (100)
205 (100)
53 (55.2)
56 (49.1)
84 (41.0)
51.3 (18.2)
49.2 (21.4)
50.0 (17.7)
17 (17.7)
18 (15.8)
22 (10.7)
68 (70.8)
72 (63.2)
140 (68.3)
Corridor Transsphenoidal Transethmoid
42 (43.8)
13 (11.4)
24 (11.7)
Transnasal
10 (10.4)
25 (21.9)
56 (27.3)
4 (6.1)
12 (10.5)
12 (5.9)
61 (63.5)
48 (42.1)
108 (52.7)
Transmaxillary Reconstruction Autologous fat graft
puncture (LP) opening pressures or radiographic evidence of skull base erosion consistent with elevated ICP. In their series of 56 patients, 82% were obese (average BMI, 36.2 kg/m2). LP opening pressures were measured on 48 of 56 patients (average, 27 cm H2O; range, 9e60 cm H2O) and lumbar drains were placed on all patients under their ICP management protocol. VP shunts or lumboperitoneal shunts were placed in 13 patients for severely elevated ICP. Of these 56 patients, 6 developed recurrent leaks (11% failure rate). Of these 6 failures, 3 patients developed leaks at new skull base sites and 2 of these patients presented with a clogged VP or lumboperitoneal shunt at the time of the leak. These 3 patients suggest a failure of ICP management rather than a failure of the original repair. Thus, regardless of the approach or technique used, successful closure of spontaneous CSF leaks requires aggressive longterm postoperative ICP management.
Gasket seal
31 (32.3)
14 (12.3)
22 (10.7)
LD preoperation
72 (75.0)
24 (21.1)
57 (27.8)
CONCLUSIONS
Postoperative CSF leak
3 (3.1)
3 (2.6)
12 (5.9)
Reoperation required
1 (1.0)
2 (1.8)
5 (2.4)
Endoscopic endonasal repair of CSF leaks and encephaloceles has evolved significantly during the past decade. Safety, efficacy, and versatility have improved dramatically and have surpassed traditional open approaches in these respects. The versatility of different endoscopic approaches through the 4 endonasal corridors allows the endoscopic repair of almost all skull-base defects. The use of vascularized pedicled mucosal flaps has evolved to cover these defects as part of multilayered closure strategies. In addition, mounting evidence suggests the superiority of these flaps—improving closure rates and healing times. As the standard of care evolves, so do our expectations. Spontaneous CSF leaks that may once have recurred at a rate that exceeded 30% are now being closed with >95% primary success. New innovations, new strategies, and new 3-dimensional technologies promise to broaden the appeal of endoscopic endonasal surgery even further.
NSF, nasoseptal flap; LD, lumbar drain; CSF, cerebrospinal fluid.
ventriculoperitoneal (VP) shunt, and use of acetazolamide. Although we place a lumbar drain after surgery for 3 days, we do not use VP shunts or acetazolamide unless the first attempt at endonasal closure fails. We place lumbar drains preoperatively on all patients undergoing CSF leak repair and on all patients with skull-base tumors who are likely to develop an intraoperative CSF leak. However, the uniform placement of a lumbar drain on all patients undergoing endoscopic CSF leak repair is controversial. Casiano and Jassir (9) reported on their series of 33 patients undergoing endoscopic CSF leak repair in 1999. In that series, 32 of 33 (97%) leaks were successfully repaired after 1 procedure without the use of a lumbar drain. However, only 6 patients in this series had noniatrogenic (spontaneous or traumatic) leak etiologies. This is similar to another series from the 1990s by Hughes et al. (17). They reported successful primary endoscopic closure of 16 of 17 (95%) CSF leaks also without the use of a lumbar drain. In that series, 9 patients had
iatrogenic leaks, 6 patients had spontaneous leaks, and 2 patients had traumatic leaks. Postoperative leak rates in these series are comparable with our more recent data (2.9% leak rate) using vascularized pedicled flaps and lumbar drains in all patients; however, there is significant variability in the literature. In their large 800-patient series, Kassam et al. (19) reported a 15.9% leak rate after primary closure, but argued that leak rates are directly proportional to the complexity and etiology of the lesion repaired. Successful closure of patients with preoperative spontaneous CSF leak and BIH is notoriously difficult and management of postoperative ICP is especially important. Postoperative treatment with acetazolamide is widely used in these patients. Acetazolamide is a diuretic that has been shown to reduce CSF production by 48% (8). Woodworth et al. (39) described a paradigm for the treatment of these patients that included 500 mg of acetazolamide twice a day postoperatively on all patients with elevated ICP as evidenced by either elevated lumbar
WORLD NEUROSURGERY 82 [6S]: S86-S94, DECEMBER 2014
REFERENCES 1. Aiden K, Guven K, Sencer S: MRI cisternography with gadolinium-containing contrast medium: its role, advantages and limitations in the investigation of rhinorrhoea. Neuroradiology 46:75-80, 2004. 2. Algin O, Hakyemez B, Gokalp G, Ozcan T, Korfali E, Parlak M: The contribution of 3D-CISS and contrast-enhanced MR cisterography in
www.WORLDNEUROSURGERY.org
S93
PEER-REVIEW REPORTS JEFFREY C. BEDROSIAN ET AL.
detecting cerebrospinal fluid leak in patient with rhinorrhoea. Br J Radiol 83:225-232, 2010. 3. Anand VK, Murali RK, Glasgold MR: Surgical decisions in the management of cerebrospinal fluid rhinorrhoea. Rhinology 33:212-218, 1995. 4. Banks CA, Palmer JN, Chiu AG, O’Malley BW Jr, Woodworth BA, Kennedy DW: Endoscopic closure of CSF rhinorrhea: 193 cases over 21 years. Otolaryngol Head Neck Surg 140:826-833, 2009. 5. Becker SS, Duncavage JA, Russell PT: Endoscopic endonasal repair of difficult-to-access cerebrospinal fluid leaks of the frontal sinus. Am J Rhinol Allergy 23:181-184, 2009. 6. Brown SM, Tabaee A, Singh A, Schwartz T, Anand VK: Three-dimensional endoscopic sinus surgery: feasibility and technical aspects. Otolaryngol Head Neck Surg 138:400-402, 2008. 7. Cappabianca P, Cavallo LM, Colao A, de Divitiis E: Surgical complications associated with the endoscopic endonasal transsphenoidal approach for pituitary adenomas. J Neurosurg 97:293-298, 2002. 8. Carrion E, Hertzog JH, Medlock MD, Hauser GJ, Dalton HJ: Use of acetazolamide to decrease cerebrospinal fluid production in chronically ventilated patients with ventriculopleural shunts. Arch Dis Child 84:68-71, 2001. 9. Casiano RR, Jassir D: Endoscopic cerebrospinal fluid rhinorrhea repair: Is a lumbar drain necessary? Otolaryngol Head Neck Surg 121:745-750, 1999.
ENDOSCOPIC CSF AND ENCEPHALOCELE REPAIR
management based on the Mayo clinic experience from 1970 through 1981. Neurosurgery 16:313-321, 1985.
30. Schick B, Ibing R, Brors D, Draf W: Long-term study of endonasal duraplasty and review of the literature. Ann Otol Rhinol Laryngol 110:142-147, 2001.
17. Hughes RGM, Jones NS, Robertson IJA: The endoscopic treatment of cerebrospinal fluid rhinorrhea: the Nottingham experience. J Laryngol Otol 111:125-128, 1997.
31. Schlosser RJ, Bolger WE: Endoscopic management of cerebrospinal fluid rhinorrhea. Otolaryngol Clin North Am 39:523-538, 2006.
18. Jakimovski D, Bonci G, Attia M, Schwartz T: Incidence and significance of intraoperative CSF leak in endoscopic pituitary surgery using intrathecal fluorescein. World Neurosurg 82:e513-e523, 2014. 19. Kassam AB, Prevedello DM, Carrau RL, Snyderman CH, Thomas A, Gardner P, Zanation A, Duz B, Stefko ST, Byers K, Horowitz MB: Endoscopic endonasal skull base surgery: analysis of complications in the authors’ initial 800 patients. J Neurosurg 114:1544-1568, 2011. 20. Kassam AB, Thomas A, Carrau RL, Snyderman CH, Vescan A, Prevedello D, Mintz A, Gardner P: Endoscopic reconstruction of the cranial base using a pedicled nasoseptal flap. Oper Neurosurg 63:544-553, 2008. 21. Kerr JT, Chu FWK, Bayles SW: Cerebrospinal fluid rhinorrhea: diagnosis and management. Otolaryngol Clin N Am 38:597-611, 2005. 22. Leng LZ, Brown S, Anand VK, Schwartz TH: “Gasket-seal” watertight closure in minimal access endoscopic cranial base surgery. Neurosurgery 62:ONE342-ONE343, 2008.
32. Schwartz TH, Fraser JF, Brown S, Tabaee A, Kacker A, Anand VK: Endoscopic cranial base surgery: classification of operative approaches. Neurosurgery 62:991-1005, 2008. 33. Selcuk H, Albayram S, Ozer H, Ulus S, Sanus GZ, Kaynar MY, Kocer N, Islak C: Intathecal gadolinium-enhanced MR cisternography in the evaluation of CSF leakage. Am J Neuroradiol 31:71-75, 2010. 34. Sen C, Triana A: Cranial chordomas: results of radical excision. Neurosurg Focus 10:E3, 2001. 35. Seth R, Rajasekaran K, Benninger MS, Batra PS: The utility of intrathecal fluorescein in cerebrospinal fluid leak repair. Otolaryngol Head Neck Surg 143:626-632, 2010. 36. Tabaee A, Anand VK, Esposito F, Schwartz ment of spontaneous the lateral sphenoid 1070-1077, 2010.
Cappabianca P, Stamm A, TH: Endoscopic managemeningoencephalocele of sinus. J Neurosurg 112:
37. Tabaee A, Anand VK, Fraser JF, Brown SM, Singh A, Schwartz TH: Three-dimensional endoscopic pituitary surgery. Neurosurgery 64 (5 Suppl 2):288-293, 2009.
10. Castelnuovo P, Dallan I, Pistochini A, Battaglia P, Locatelli D, Bignami M: Endonasal endoscopic repair of Sternberg’s canal cerebrospinal fluid leaks. Laryngoscope 117:345-349, 2007.
23. Mattox DE, Kennedy DW: Endoscopic management of cerebrospinal fluid leaks and cephaloceles. Laryngoscope 100:857-862, 1990.
38. Tabaee A, Placantonakis DG, Schwartz TH, Anand VK: Intrathecal fluorescein in endoscopic skull base surgery. Otolaryngol Head Neck Surg 137:316-320, 2007.
11. Dehdashti AR, Ganna A, Karabatsou K, Gentili F: Pure endoscopic endonasal approach for pituitary adenomas: early surgical results in 200 patients and comparison with previous microsurgical series. Neurosurgery 62:1006-1017, 2008.
24. McCoul ED, Anand VK, Singh A, Nyquist GG, Schaberg MR, Schwartz TH: Long-term outcomes after endoscopic skull base reconstruction with the nasoseptal flap. World Neurosurg 81:136-143, 2014.
39. Woodworth BA, Prince A, Chiu AG, Cohen NA, Schlosser RJ, Bolger WE, Kennedy DW, Palmer JN: Spontaneous CSF leaks: a paradigm for definitive repair and management of intracranial hypertension. Otolaryngol Head Neck Surg 138:715-720, 2008.
25. Nyquist GG, Anand VK, Mehra S, Kacker A, Schwartz TH: Endoscopic endonasal repair of anterior skull base non-traumatic cerebrospinal fluid leaks, meningoceles and encephaloceles. J Neurosurg 113:961-966, 2010.
40. Yang Z, Wang B, Wang C, Liu P: Primary spontaneous cerebrospinal fluid rhinorrhea: a symptom of idiopathic intracranial hypertension? J Neurosurg 115:165-170, 2011.
12. Feiz-Erfan I, Han PP, Spetzler RF, Horn EM, Klopfenstein JD, Porter RW, Ferreira MA, Beals SP, Lettieri SC, Joganic EF: The radical transbasal approach for resection of anterior and midline skull base lesions. J Neurosurg 103:485-490, 2005. 13. Fraser JF, Allen B, Anand VK, Schwartz TH: Threedimensional neurostereoendoscopy: subjective and objective comparison to 2D. Minim Invasive Neurosurg 52:25-31, 2009. 14. Gassner HG, Ponikau JU, Sherris DA, Kern EB: CSF rhinorrhea: 95 consecutive surgical cases with long term follow-up at the Mayo Clinic. Am J Rhinol 13:439-447, 1999.
26. Nyquist GG, Anand VK, Singh A, Schwartz TH: Janus flap: bilateral nasoseptal flaps for anterior skull base reconstruction. Otolaryngol Head Neck Surg 142:327-331, 2010. 27. Ommaya AK, di Chiro G, Baldwin M, Pennybacker JB: Non-traumatic cerebrospinal fluid rhinorrhea. J Neurol Neurosurg Psychiatry 31: 214-225, 1968.
15. Hadad G, Bassagasteguy L, Carrau RL, Mataza JC, Kassam A, Snyderman CH, Mintz A: A novel reconstructive technique after endoscopic expanded endonasal approaches: vascular pedicle nasoseptal flap. Laryngoscope 116:1882-1886, 2006.
28. Placantonakis DG, Tabaee A, Anand VK, Schwartz TH: Safety of low-dose intrathecal fluorescein in endoscopic cranial base surgery. Oper Neurosurg 61:161-166, 2007.
16. Hubbard JL, McDonald TJ, Pearson BW, Laws ER Jr: Spontaneous cerebrospinal fluid rhinorrhea: evolving concepts in diagnosis and surgical
29. Sautter NB, Batra PS, Citardi MJ: Endoscopic management of sphenoid sinus cerebrospinal fluid leaks. Ann Otol Rhinol Laryngol 117:32-39, 2008.
S94
www.SCIENCEDIRECT.com
41. Zuckerman JD, DelGaudio JM: Utility of pre-operative high-resolution CT and intraoperative image guidance in identification of cerebrospinal fluid leaks for endoscopic repair. Am J Rhinol 22:151-154, 2008.
Conflict of interest statement: The authors declare that the article content was composed in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Received 18 September 2013; accepted 24 July 2014 Citation: World Neurosurg. (2014) 82, 6S:S86-S94. http://dx.doi.org/10.1016/j.wneu.2014.07.018 Journal homepage: www.WORLDNEUROSURGERY.org Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2014 Elsevier Inc. All rights reserved.
WORLD NEUROSURGERY, http://dx.doi.org/10.1016/j.wneu.2014.07.018
The Endoscopic Endonasal Approach to Repair of Iatrogenic and Noniatrogenic Cerebrospinal Fluid Leaks and Encephaloceles of the Anterior Cranial Fossa Jeffrey C. Bedrosian2, Vijay K. Anand2, Theodore H. Schwartz1,3
Key words Anterior skull base - Endoscopic - Minimal access - Minimally invasive - Spontaneous CSF leak - Traumatic CSF leak -
Abbreviations and Acronyms BIH: Benign intracranial hypertension BMI: Body mass index CSF: Cerebrospinal fluid CT: Computerized tomography FESS: Functional endoscopic sinus surgery HRCT: High resolution CT ICP: Intracranial pressure MRI: Magnetic resonance imaging VP: Ventriculoperitoneal Departments of 1Neurological Surgery, 2 Otolaryngology—Head and Neck Surgery, 3 Neurology and Neuroscience, Weill Cornell Medical College, New York Presbyterian Hospital, New York, New York, USA To whom correspondence should be addressed: Theodore H. Schwartz, M.D. [E-mail: [email protected]] Citation: World Neurosurg. (2014) 82, 6S:S86-S94. http://dx.doi.org/10.1016/j.wneu.2014.07.018 Journal homepage: www.WORLDNEUROSURGERY.org
- OBJECTIVE:
The current approach for the diagnosis and repair of spontaneous and traumatic anterior skull-base defects is oulined, highlighting the controversies that exist in the field and describing the strategies required to access different segments of the anterior cranial fossa.
- METHODS:
We reviewed the literature concerning endoscopic management of anterior skull-base defects. These publications have been combined with our own experience repairing cerebrospinal fluid (CSF) leaks and encephaloceles that developed spontaneously, traumatically, or intentionally as a result of endoscopic skull-base surgery.
- RESULTS:
We present a systematic methodology for the repair of these defects. We have divided our surgical approach into four separate corridors. These are the transnasal, transsphenoidal, transethmoidal, and transmaxillary corridors. Dissection strategies vary for each corridor, but with a combination of approaches, all areas of the anterior skull base can be accessed. Skull-base defects are successfully repaired with a multilayered closure that often involves use of a vascularized pedicled mucosal flap. Adoption of this technique has decreased our rate of postoperative CSF leak from 5.9%e3.1%.
- CONCLUSIONS:
Endoscopic endonasal repair of CSF leaks and encephaloceles has evolved significantly during the past decade. The versatility of different endoscopic approaches through the four endonasal corridors allows for the endoscopic repair of almost all skull-base defects. The use of vascularized pedicled mucosal flaps has evolved to cover these defects as part of multilayered closure strategies.
Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2014 Elsevier Inc. All rights reserved.
INTRODUCTION Noniatrogenic cerebrospinal fluid (CSF) leaks occur commonly in the anterior cranial fossa. Iatrogenic CSF leaks can occur inadvertently during endoscopic sinus surgery and as a natural course of endoscopic skull-base surgery. For spontaneous CSF leaks, primary closure rates have improved from 70%e80% in the 1980s (16) to >90% with the evolution of endoscopic techniques (14, 23, 25, 30). As such, endoscopic endonasal repair of CSF leaks and encephaloceles has begun to replace traditional open craniotomy techniques. Closure rates for iatrogenic high-flow CSF leaks that occur during endoscopic skull-base surgery have improved from 20%e70% in early publications to 25 cm H2O and normal biochemical and cytologic composition of CSF 6. No other explanation for the increased intracranial pressure CT, computerized tomography; MRI, magnetic resonance imaging; LP, lumbar puncture; CSF, cerebrospinal fluid.
lateral defects may occur due to improper fusion of Sternberg’s canal, a phenomenon discussed later. In addition, these encephaloceles tend to be quite large, as increased ICP pushes intracranial contents
ENDOSCOPIC CSF AND ENCEPHALOCELE REPAIR
through relatively small skull-base defects. Spontaneous CSF leaks are also highly associated with the female gender and obesity. In one series of 55 consecutive patients with spontaneous CSF leaks, 70% of patients were women and 46 of 55 patients were obese, with an average body mass index (BMI) of 36.2 kg/m2. Elevated ICP persisted in these patients postoperatively, with an average lumbar drain pressure of 27 cm H2O (39). In another series of 21 patients, 18 patients were women, with an average BMI of 31.2 kg/m2 (40). In many ways, spontaneous CSF leaks are the most difficult to treat. Small, slow leaks can be hard to locate. Broad attenuation of the bone of the skull base makes achieving a watertight closure difficult. Associated encephaloceles or meningoencephaloceles are found in 50%e100% of patients. Postoperative recurrence rates are generally higher compared with all other causes, ranging from 25%e87% in some series (16, 27, 30). Important, understanding the etiology and demographics of spontaneous CSF leaks is critical to
successful primary closure. Although repair techniques are similar to the closure of other leak types, postoperative ICP management of these patients is critical. The details and controversies of postoperative ICP management will be discussed later. Traumatic CSF leaks may be divided into iatrogenic and noniatrogenic leaks. Iatrogenic CSF leaks most commonly occur during transsphenoidal pituitary tumor resection (0.5%e15% incidence). They also frequently occur during acoustic neuroma surgery (7%e11% incidence) and during functional endoscopic sinus surgery (0.5%e3% incidence) (21). The most common sites of inadvertent skull base violation during endoscopic sinus surgery are the lateral lamella of the cribriform plate and the roof of the posterior ethmoid sinuses (3), as the skull base slopes inferiorly toward the face of the sphenoid sinuses (Figure 2). Of these leaks, 50% present intraoperatively or immediately postoperatively, whereas the remaining 50% are delayed, occurring 1 week to 1 month postoperatively. Theories for delayed iatrogenic CSF leak presentation include wound contraction, flap devascularization, or necrosis, resolving cerebral edema and increased ICP (21). Traumatic noniatrogenic CSF leaks are usually due to accidental trauma (70%e 80% incidence). Two to 4% of all acute head injuries result in CSF rhinorrhea. In 70% of cases, these leaks resolve spontaneously with observation or a lumbar drainage; however, repair should be considered if leaks do not resolve within 1e2 weeks to prevent meningitis (21). Intracranial tumors may cause CSF leaks as well. Tumor growth may obstruct CSF reabsorption leading to increased ICP and a high pressure leak. Alternatively, intracranial or extracranial tumor growth may directly erode the skull base. LOCALIZING THE DEFECT
Figure 1. Artist’s illustration of the fovea ethmoidalis (inset), a frequent site for inadvertent skull base violation. A sagittal view of the skull base demonstrating possible skull base defect locations and encephalocele formation. Defects shown are at the foramen cecum (anterior) and the posterior cribriform region (posterior) as the skull base slopes downward to meet the face of the sphenoid sinus.
WORLD NEUROSURGERY 82 [6S]: S86-S94, DECEMBER 2014
Preoperative and intraoperative localization of a CSF leak is an essential component of successful repair. There are several options for preoperative localization. High resolution CT (HRCT), CT cisternography, magnetic resonance imaging (MRI), MRI cisternography, and radionuclide cisternography will be discussed. HRCT is ideally noninvasive; it does not require any
www.WORLDNEUROSURGERY.org
S87
PEER-REVIEW REPORTS JEFFREY C. BEDROSIAN ET AL.
ENDOSCOPIC CSF AND ENCEPHALOCELE REPAIR
Figure 2. An iatrogenic defect of the fovea ethmoidalis (crosshairs) is shown. A small encephalocele is present.
intrathecal injection. HRCT is capable of resolving millimeter-sized skull-base defects and as such, its chief drawback is the detection of congenital skull-base defects not associated with CSF leak or encephalocele. HRCT also exposes patients to relatively high doses of radiation. In 1 study, high resolution CT was used in 40 patients with anterior skull-base defects. Forty-two defects were detected with HRCT. The investigators correlated their findings with intraoperative image guidance, achieving a primary endoscopic endonasal repair rate of 95% (40/42 defects). This rate improved to 100% after revision surgery (41). CT cisternography requires injection of contrast material intrathecally with a LP. This technique is 85% sensitive for detecting active leaks at the time of the study, but its sensitivity varies between 48% and 96% for inactive leaks. Contrast material does not distribute evenly throughout the CSF and the small quantity
S88
www.SCIENCEDIRECT.com
of contrast passing through a hairline leak may not be able to be resolved from the surrounding bone (33). CT cisternography is useful when the frontal sinus or sphenoid sinus serves as a reservoir for accumulating CSF fluid, as these reservoirs will demonstrate contrast enhancement (21). The drawbacks of CT cisternography are its radiation exposure, especially if delayed scans are required, its invasive nature, possible contrast allergic reactions, and rarely intracranial hemorrhage. T2-weighted MRI has been used to detect high signal intensity fluid communicating between the sub-arachnoid space and the paranasal sinuses or encephalocele herniation into the sinuses. This technique lacks specificity, however, and may be confused with preexisting paranasal sinus inflammation, leading to an unacceptably high false-positive rate. MRI cisternography requires gadolinium injection into the subarachnoid space. In a prospective study of 20 patients with both
active and intermittent CSF rhinorrhea, a CSF leak was localized in 16 patients after MRI cisternography with gadolinium. T1 fat-suppressed images were taken before and after injection in a 30- to 40-degree prone head-down position during a 15- to 30-minute period. Aiden et al. (1) reported no adverse side effects from the 0.5-mL intrathecal gadolinium injection. This technique, although not yet widely adopted, offers a promising alternative to HRCT or CT cisternography and their relatively high levels of radiation exposure. MRI imaging is also useful for the differentiation of meningoceles and meningoencephaloceles. In another study using MRI cisternography, Algin et al. (2) reported a 100% sensitivity for detection of CSF leakage. In a larger study of 85 patients, MRI cisternography detected a CSF leak in 100% of patients with clinically evident meningitis or continuous CSF rhinorrhea (33). CSF leak was detected in 70% of patients with confirmed intermittent leakage and 60% of patients with suspected CSF rhinorrhea (33). Radionuclide cisternography may be conducted during several days and is ideal for low volume intermittent leaks. It uses technetium 99 administered intrathecally. Its chief drawback is its poor spatial resolution compared with CT and MRI cisternography. Intranasal pledgets may be placed endoscopically to help localize the leak. Radionuclide cisternography has a high 33% false-positive rate, with a sensitivity that varies between 62% and 76%. For intraoperative localization, one can use intrathecal fluorescein. The safety of routine use of intrathecal fluorescein, an off-label application of the drug, has been debated in the literature. Our protocol for administering intrathecal fluorescein is as follows: informed consent for intrathecal fluorescein injection is obtained. After intubation, 10 mg of dexamethasone and 50 mg of diphenhydramine are administered to the patient. A LP is performed, with or without placement of a lumbar drain. A total of 10 mL of CSF is withdrawn and mixed with 0.25 mL of 10% fluorescein. The CSF/fluorescein mixture is slowly reinjected into the intrathecal space. In our own retrospective review of 54 patients undergoing endoscopic skull-base surgery, 3 of 54 patients experienced postoperative lower extremity weakness or numbness. These results were possibly confounded by
WORLD NEUROSURGERY, http://dx.doi.org/10.1016/j.wneu.2014.07.018
PEER-REVIEW REPORTS JEFFREY C. BEDROSIAN ET AL.
placement of a lumbar drain in 2 of the 3 patients (28). In a larger series (18) of 203 pituitary adenomas, there were no clear fluorescein-related complications. Intrathecal fluorescein is an invaluable tool for detection of intraoperative CSF leak and is highly predictive of postoperative CSF leak. In a retrospective analysis (38) of 61 patients undergoing endoscopic skull-base surgery for a variety of lesions, 7 of 37 patients (19.4%) with an intraoperative fluorescein leak experienced postoperative CSF leak. Of the remaining 24 patients, no fluorescein leak was visualized. There was a 0% incidence of postoperative leak among these patients (38). With respect to repair of preexisting skull-base defects, identification of intraoperative fluorescein leak is essential. For small leaks, use of a blue light filter attached to the endoscope greatly enhances the sensitivity of detection of fluorescein. With this technique, fluorescein may be visually resolved at 1 part in 10 million (21). In a retrospective case series of 103 patients undergoing CSF leak repair, intrathecal fluorescein was used in 47 cases. Most were cases of spontaneous CSF leak (61.7%). Of these patients, fluorescein was visualized at the leak site in 66% of cases. A leak was identified without visualization of fluorescein in 23.4% of cases (false negatives). In 10.6% of cases (5 patients), no leak was identified and no fluorescein was identified. Nonetheless, the most likely site of leak was repaired. Of these 5 patients, 2 developed postoperative recurrent CSF leak. Fluorescein sensitivity for detecting an intraoperative CSF leak was 73.8% and specificity was 100%. The rate of recurrent leak was 31.3% when fluorescein was not visualized compared with 9.7% when fluorescein was visualized; however, this finding was not statistically significant (P ¼ 0.10) (35). ADVANTAGES AND DISADVANTAGES OF THE ENDOSCOPIC ENDONASAL APPROACH Endoscopic, endonasal skull-base approaches have significant advantages when compared with traditional transcranial open surgery. Endoscopic approaches do not require brain retraction or neurovascular manipulation. They do not require craniotomies and large scalp incisions. They do not result in a cosmetic deformity. As a result, the complication
ENDOSCOPIC CSF AND ENCEPHALOCELE REPAIR
rate and hospital stay after these surgeries is significantly decreased. In addition, when compared with the microscope, the endoscope offers a wider, more panoramic view of the surgical field. This is due to the position of the endoscope lens in close proximity to the operative site, versus a microscope lens several centimeters more distant. By the same principle, the endoscope offers superior lighting of the operative field compared with the microscope. Endoscopes can also have angled lenses on their tips, providing the ability to look around corners and operate in areas obscured from the line-of-site limitations of the microscopic view. Advances in video screen and camera technology allow the endoscopic surgeon to enjoy a high-definition view of the surgical field that is comparable to the clarity observed through the oculars of a microscope. The principle drawback of endoscopic surgery is the lack of binocular vision provided by the microscope. Otolaryngologists tend to be familiar and comfortable with this 2-dimensional endoscopic picture because of their extensive experience with endoscopic sinus surgery. Indeed, 3 dimensionality can be approximated by continued movement of the endoscope in and out of the nose, as surgical landmarks change size on the screen relative to their proximity to the endoscope lens. For neurosurgeons used to working through the microscope, performing microsurgery in a 2-dimensional world requires acclimation and is associated with a learning curve. Within the past few years, however, 3dimensional endoscopes have become commercially available. Although not widely used, these new endoscopes are capable of displaying the 3-dimensional contours of the skull base, sinonasal cavity, and sella (6, 13, 37). At present, these endoscopes do not offer the same highdefinition resolution as the standard endoscopic view, and they require the surgeon to wear 3-dimensional glasses. Nonetheless, as the technology improves, use of 3-dimensional endoscopes will undoubtedly become more widespread, with visual clarity and depth of field approaching the microscope’s standard. ENDONASAL OPERATIVE APPROACHES CSF leaks can occur anywhere throughout the anterior cranial fossa. As such, a
WORLD NEUROSURGERY 82 [6S]: S86-S94, DECEMBER 2014
variety of endoscopic approaches are necessary to access these disparate regions of the skull base. In general, the degree of dissection and degree of difficulty required increases as one moves further away from the midline; however, the basic principles of successful endoscopic closure always apply. We have previously classified our broad array of endoscopic approaches for resection of skull-base tumors based on 1) an anatomic target, 2) a cranial base approach, and 3) a nasal corridor. This scheme can also be applied to strategies for repair of CSF leaks and encephaloceles (32). We identified 12 separate anatomic skullbase targets. Of those, the ones likely associated with a CSF leak include the anterior fossa, olfactory groove, sella, suprasellar cistern, and pterygopalatine fossa. These regions can be approached through four possible corridors: transnasal, transsphenoidal, transethmoidal, and transmaxillary (Figure 3). Understanding these corridors anatomically is necessary for appropriate and safe dissection performed by the otolaryngologist during the approach to the skull-base defect target. In this context, we define the endonasal approach as a surgical strategy based on the nasal corridor traversed on our way to the target. Therefore, it is the specific dissection technique that gets us from the nostril to the target. Each approach applicable to the repair of a potential CSF leak will be described in this manner. The transcribriform approach is the most midline approach. Access to the cribriform plate is gained through the transnasal corridor. The transnasal corridor is defined as those spaces within the nose and skull base accessed without transgression of the sinuses. Strictly speaking, all endonasal procedures begin in the transnasal corridor, but midline structures like the olfactory groove and cribriform plate may be accessed without further dissection. The lateral boundaries of this approach are the middle turbinates bilaterally. Encephaloceles and meningoencephaloceles often herniate into this region of the nose. Often it is necessary to perform a partial septectomy, removing the perpendicular plate of the ethmoid bone to circumscribe these lesions by operating through both nostrils. Doing so, however, will invariably damage the olfactory mucosa, which extends inferiorly from
www.WORLDNEUROSURGERY.org
S89
PEER-REVIEW REPORTS JEFFREY C. BEDROSIAN ET AL.
ENDOSCOPIC CSF AND ENCEPHALOCELE REPAIR
Figure 3. An artist’s illustration of the four sinonasal corridors with the skull in three-quarter projection.
the cribriform plate onto the superior turbinates. This leads to anosmia. This basic approach exposes a narrow segment of the skull base centered on the crista galli. When more lateral access is required, a transfovea ethmoidalis approach is required. The fovea ethmoidalis is a common site of accidental skull base transgression during endoscopic sinus surgery (Figure 4). The foveal ethmoidalis is accessed through the transethmoidal corridor. In this corridor, the middle turbinates are now the medial boundaries
Figure 4. The fovea ethmoidalis is a frequently violated anatomic landmark during endoscopic sinus surgery (arrow).
S90
www.SCIENCEDIRECT.com
and the lamina paprecea of the orbits are the lateral boundaries of the dissection. Otolaryngologists are used to developing the transethmoidal corridor during a complete ethmoidectomy, a routine element of functional endoscopic sinus surgery (FESS). This requires removal of the uncinate process, a thin vertically oriented bone that visually obscures the natural ostium of the maxillary sinus and the ethmoid bulla. The ethmoid bulla is the anterior-most ethmoid air cell and its removal is the first step in a complete ethmoidectomy. Identification of the skull base so that it can be avoided is of primary importance during this operation. Nonetheless, the lateral lamella of the fovea is the thinnest bone of the skull base and is easily violated. Such iatrogenic defects are usually noticed at the time of sinus surgery. If such an injury occurs, immediate repair is advised. In this case, the surgical corridor is already present as part of the dissection for the FESS. Development of this corridor in combination with the transnasal corridor provides lamina to the lamina skull base access in the coronal plane. It also provides access from the
bony beak of the frontal recess to the face of the sphenoid sinus in the sagittal plane. CSF leaks may be in the sphenoid sinus. They can occur as a result of iatrogenic accidental trauma, or intentionally, after removal of sellar, suprasellar, or cavernous sinus lesions. They can occur spontaneously as a result of incompetence of the diaphragma sellae with meningoencephalocele herniation leading to primary empty sella, or even empty sella syndrome. Alternatively, far lateral sphenoid sinus leaks may occur due to congenital skull base dehiscence in this region. In all patients, the transsphenoidal corridor provides access to the sphenoid sinus. The anatomic targets reached through the transsphenoidal corridor are the sella and the suprasellar sistern, accessed with the transsellar and transtuberculum approach, respectively. The cavernous sinus and planum sphenoidale may be accessed with these approaches as well. The endonasal corridor provides access to the face of the sphenoid sinus. The natural ostia of the sphenoid sinuses can be located by directing the endoscope medially to the middle turbinate. These ostia are then widely enlarged bilaterally. If one wishes to work in the sphenoid sinuses bimanually, a posterior septectomy can be performed with or without preservation of a nasoseptal flap (discussed later). The intersinus septum can then be drilled away allowing access to the breadth of the sphenoid sinus by both the endoscope and surgical instruments necessary to repair the leak. If the leak is located in the suprasellar region, the planum sphenoidale and tuberculum sella may be drilled away once the transsellar corridor has been created. To achieve adequate visualization and instrumentation of all these transsphenoidal approaches, it is essential to widely open the face of the sphenoid sinus bilaterally. This may require removal of both superior turbinates. The intersinus septum and any partial sphenoid septations should be drilled flush with the face of the sella and opticocarotid recess with a diamondtipped burr. Only after the completion of these maneuvers can the entire height and breadth of the sphenoid sinus be appreciated (32). Lateral sphenoid sinus CSF leaks may occur as a result of a dehiscence of Sternberg’s canal. During development,
WORLD NEUROSURGERY, http://dx.doi.org/10.1016/j.wneu.2014.07.018
PEER-REVIEW REPORTS JEFFREY C. BEDROSIAN ET AL.
each half of the sphenoid bone requires the fusion of five cartilaginous precursors. Normal development of the greater wing of the sphenoid requires complete fusion with the presphenoid and basisphenoid precursors. When this does not occur, a lateral craniopharyngeal canal (Sternberg’s canal) remains (Figure 5). This skull base dehiscence occurs in the lateral recess of the sphenoid sinus and is not adequately visualized through the transsphenoidal corridor (36). A CSF leak or meningoencephalocele has been reported to occur through this dehiscence (Figure 6). At present, several case series (10, 29, 36) have been published addressing the treatment of this rare clinical entity. Endoscopic access to the far lateral sphenoid requires a combination of the transsphenoidal, transethmoidal, and transpterygoid approaches. The transpterygoid approach is accessed by the transmaxillary corridor. The transmaxillary corridor is developed by identification of the natural ostium of the maxillary sinus after removal of the uncinate process. The maxillary ostium can then be widely enlarged using standard FESS techniques to visualize the posterior wall of the maxillary sinus. The posterior wall of the maxillary sinus may then be drilled away, providing access to the pterygopalatine fossa and its neurovascular contents. The sphenopalatine artery is then identified and ligated. The remaining neurovascular structures of the pterygopalatine fossa can then be reflected laterally so that the pterygoid process may be drilled down. This maneuver provides access to the lateral recess of the sphenoid sinus, revealing the origin of Sternberg’s canal. When this approach is combined with the transethmoidal, transsphenoidal approaches, the sphenoid sinus can be widely visualized and instrumented. Using
Figure 5. An artist’s illustration of the sphenoid bone. The arrow points to a lateral dehiscence of Sternberg’s canal.
ENDOSCOPIC CSF AND ENCEPHALOCELE REPAIR
this technique, Tabaee et al. (36) presented a case series of 13 patients with far lateral meningoencephaloceles. In their series 11 of 13 patients were successfully closed after 1 surgical attempt. Two patients developed postoperative CSF rhinorrhea. Of these, 1 postoperative leak resolved spontaneously and the other required reoperation. All patients were followed for an average of 4.7 years ( 3.3 years) without leak recurrence. Facial numbness is a potential morbidity of this approach, given the proximity of the infraorbital nerve within the pterygopalatine foramen. In their series, Tabaee et al. (36) reported 1 of 13 patients with postoperative facial paresthesia. There were no reported complications in a series of 15 patients by Castelnuovo et al. (10). THE ENDOSCOPIC ENDONASAL APPROACH TO THE FRONTAL SINUSES The frontal sinuses are the most difficult area of the skull base to reach endoscopically. CSF leaks and encephaloceles of the frontal sinuses have traditionally been repaired through a pure open approach with the creation of an osteoplastic flap, or through combined endoscopic and open approaches. It may be difficult to visualize and repair CSF leaks of the far lateral frontal sinus through endoscopic means alone. Even if a defect can be visualized, a complex multilayered repair is usually not possible. In addition, the narrow frontal sinus outflow tract makes placement of a vascularized pedicled mucosal flap impossible. In their series of 28 consecutive patients with 32 various skull-base leaks, Nyquist et al. (25) reported only 2 failures after endoscopic repair. One of these failures, however, was located in the posterior wall of the frontal sinuses. This recurrence, which occurred after a vehicular trauma, was identified adjacent to a previously repaired large meningoencephalocele defect site. A transcranial approach for the revision surgery was chosen due to the risk of frontal sinus outflow tract stenosis. Becker et al. (5) reported 2 cases of purely endoscopic endonasal repair of frontal sinus CSF leaks. One patient developed a CSF leak of the lateral frontal sinus. This lesion was visualized with a frontal sinus drillout (Draf III/Modified Lothrop procedure) and creation of a
WORLD NEUROSURGERY 82 [6S]: S86-S94, DECEMBER 2014
Figure 6. The lateral dehiscense of Sternberg’s canal is usually not visualize through the sphenoid sinus. Instead a transpterygoid fossa approach is required to identify the encephalocele.
25 35 mm frontal sinusotomy. A fat graft was placed to seal the defect; however, the leak recurred 19 days later. This recurrence was successfully treated with lumbar drain placement for 7 days. This group also reported successful primary repair of a 5 7 mm superior frontal sinus defect using a similar technique (5). ENDOSCOPIC ENDONASAL CLOSURE TECHNIQUES Spontaneous CSF Leaks Various graft materials and multilayer closure strategies have been successfully used to seal skull-base defects; however, there are several basic principles that are universally applicable. First, the defect must be circumferentially visualized. If an encephalocele or meningoencephalocele is present, it should be reduced or cauterized flush with the skull base, as herniated brain tissue is typically nonfunctional. Next, the skull base should be circumferentially demucosalized around the defect site. There should be no overlap between the in situ skull-base mucosa and the mucosal graft applied during the closure. Any such overlap may result in development of a mucocele that will expand with time and require a further operation to correct. Finally, it is critical that the mucosal overlay graft lie flush with the skull base. Occasionally, the defect may need to be recontoured so that it is flat enough to allow the graft to make adequate contact with the skull base (4, 31).
www.WORLDNEUROSURGERY.org
S91
PEER-REVIEW REPORTS JEFFREY C. BEDROSIAN ET AL.
Grafting may be performed with an underlay technique, with an overlay technique, or with a combination of both techniques. Underlay grafting require placement of a graft between the dura and the skull base, resting against the cranial side of the skull base. Options for underlay graft materials include septal mucoperiosteum, temporalis fascia, Duragen, cadaveric fascia, dermis, or pericardium. Closure rates are not affected by the type of underlay used (31). All grafts should circumferentially overlap all edges of the defect, resting flush with the cranial surface of the skull base as part of a multilayered closure. Placement of an underlay graft for defects smaller than 4 mm may be difficult. In these cases, overlay grafting alone is usually sufficient. We often interpose a fat graft between these small defects and our mucosal graft, sealing the multilayer closure in place with Duraseal. If overlay grafting is needed, it may be performed with either a free mucosal graft or a vascularized pedicled mucosal flap. Free mucosal grafts are well-described in the literature. Sources for these grafts often come from adjacent septal or inferior turbinate mucosa. Vascularized pedicled mucosal flaps may be derived from the nasoseptal mucosa, the middle turbinate mucosa, the inferior turbinate mucosa, or the nasal floor mucosa. Choice of the source of the flap depends on the size and location of the defect to be covered. The nasoseptal flap is the most popular vascularized pedicled flap (15). It is the easiest to harvest, has a robust blood supply, and offers the largest coverage and the longest pedicle. This flap is based on the nasoseptal artery, a terminal branch of the sphenopalatine artery. The pedicle for the flap spans the distance between the natural ostium of the sphenoid sinus and the ipsilateral apex of the choana. The septal mucosa can be harvested quite anteriorly, stopping at a point that approximates the anterior tip of the inferior turbinate (20). For very large skull-base defects, two nasoseptal flaps can be harvested from each side of the nasal septum. This technique has been termed the Janus flap as the two flap faces may be onset adjacent to each other to cover a large anteroposterior defect (26). Iatrogenic CSF Leaks Transsphenoidal excision of pituitary macroadenomas accounts for about half of the
S92
www.SCIENCEDIRECT.com
ENDOSCOPIC CSF AND ENCEPHALOCELE REPAIR
skull-base tumors removed in our practice. In our experience, the rate of iatrogenically created intraoperative CSF leak rises as tumor size increases (18). As previously described, intraoperative CSF leak is easily verified by preoperative injection of intrathecal fluorescein. After tumor removal, if a CSF leak is present, the sellar defect is closed in a multilayered fashion. After hemostasis is achieved, a free fat graft, usually taken from the abdomen, is placed in the defect. Care is taken not to over-pack the fat in the defect to prevent compression of the normal pituitary. Medpore is then sized to be slightly larger than the bony sellar defect and inset into the sella to support the fat graft. An overlay vascularized pedicled flap is then applied and the closure is sealed with Duraseal. Adoption of the vascularized pedicled flap for the multilayered closure of sellar defects decreased our postoperative CSF leak rate from 5.9%e2.9% in our own series of 135 patients (Table 2) (24). For large skull-base defects that are intentionally created during the removal of nonpituitary skull-base tumors, we have developed a gasket seal closure technique (22). After circumferentially defining the defect, a fascia lata graft is harvested. The dimensions of this graft are at least 1 cm larger on all sides than the size of the defect. This allows the graft to be partially inset into the defect. The graft is held in place with Medpore, sized to rest on the cranial surface of the skull base in an underlay fashion. In this way, the fascia lata creates a watertight seal, as parts of the graft lie both intracranially and extracranially. Intracranial pressure holds this gasket seal tightly against the skull base and a vascularized pedicled mucosal graft completes the closure in an overlay fashion (22). We analyzed our own series of 415 skull base cases to determine the benefit of the nasoseptal flap for the prevention of postoperative CSF leak. Of the 301 cases, 205 were performed prior to adoption of the nasoseptal flap and 210 cases were performed after the adoption of the nasoseptal flap in March 2008. Of these 210 patients, 96 cases had a nasoseptal flap (Table 2). All four endonasal corridors were represented in our series. After March 2008, our rate of postoperative CSF leak decreased by approximately one-half, from 5.9% (before March 2008) to 3.1% (after March 2008 nasoseptal flap group) and 2.6% (after March 2008 non-nasoseptal
flap group). Our improved non-nasoseptal flap closure rate after March 2008 can be explained by more selective use of this technique for smaller skull-base defects, usually without an intraoperative CSF leak. POSTOPERATIVE MANAGEMENT Postoperative antibiotic use to prevent meningitis is controversial. We use antibiotics with Gram-positive coverage for as long as nasal packing remains. However, nasal packing is typically removed within the first 24 hours postoperatively. Telfa nasal packing placed in both nares is used to minimize postoperative nasal bleeding and also to bolster our closure while the Duraseal placed over the mucosal flap hardens. For nonpituitary skull-base tumors, triple antibiotic coverage with vancomycin, metronidazole, and a second or third generation cephalosporin is used for the first 72 hours. Despite the nonsterile nature of the nose, postoperative intracranial infection is rare. In a comprehensive review of their experience with 800 skull base patients, Kassam et al. (19) reported a 1.9% rate of postoperative infection confirmed by bacterial culture. Among these, 13 cases resulted in meningitis, 1 case resulted in an intradural abscess, and 1 case resulted in an extradural abscess. For purely endoscopic pituitary resections, the rate of postoperative infection has been reported to be even lower (0.7%e1.0%) in two large series (7, 11). These postoperative infection rates are significantly decreased when compared with traditional open approaches (12, 34) despite the nonsterile nature of the nose. We speculate that the low rate of postoperative infection is due to the relatively small skull base opening of the endoscopic approach when compared with traditional open craniotomies. Postoperative management of elevated ICP in patients with high pressure tumorrelated leaks or spontaneous leaks with BIH can help prevent leak recurrence (39). After surgery, the head of bed is elevated to 30 degrees. Patients are placed on stool softeners and instructed to avoid nose blowing, drinking through a straw, sneezing with their mouth closed, or lifting objects heavier than 5 pounds for the next 2 weeks. Other adjunctive measures for the management of postoperative elevated ICP can include placement of a lumbar drain, placement of a
WORLD NEUROSURGERY, http://dx.doi.org/10.1016/j.wneu.2014.07.018
PEER-REVIEW REPORTS JEFFREY C. BEDROSIAN ET AL.
ENDOSCOPIC CSF AND ENCEPHALOCELE REPAIR
Table 2. Our Series of 415 Skull Base Patients Before and After the Adoption of the Nasoseptal Flap After Adoption of NSF Reconstruction in March 2008
Total number of patients Men Age, mean (SD) (years) Prior surgery
Group A (NSF), (n, %)
Group B (non-NSF), (n, %)
Before Adoption of NSF Reconstruction in March 2008 (group C), (n, %)
96 (100)
114 (100)
205 (100)
53 (55.2)
56 (49.1)
84 (41.0)
51.3 (18.2)
49.2 (21.4)
50.0 (17.7)
17 (17.7)
18 (15.8)
22 (10.7)
68 (70.8)
72 (63.2)
140 (68.3)
Corridor Transsphenoidal Transethmoid
42 (43.8)
13 (11.4)
24 (11.7)
Transnasal
10 (10.4)
25 (21.9)
56 (27.3)
4 (6.1)
12 (10.5)
12 (5.9)
61 (63.5)
48 (42.1)
108 (52.7)
Transmaxillary Reconstruction Autologous fat graft
puncture (LP) opening pressures or radiographic evidence of skull base erosion consistent with elevated ICP. In their series of 56 patients, 82% were obese (average BMI, 36.2 kg/m2). LP opening pressures were measured on 48 of 56 patients (average, 27 cm H2O; range, 9e60 cm H2O) and lumbar drains were placed on all patients under their ICP management protocol. VP shunts or lumboperitoneal shunts were placed in 13 patients for severely elevated ICP. Of these 56 patients, 6 developed recurrent leaks (11% failure rate). Of these 6 failures, 3 patients developed leaks at new skull base sites and 2 of these patients presented with a clogged VP or lumboperitoneal shunt at the time of the leak. These 3 patients suggest a failure of ICP management rather than a failure of the original repair. Thus, regardless of the approach or technique used, successful closure of spontaneous CSF leaks requires aggressive longterm postoperative ICP management.
Gasket seal
31 (32.3)
14 (12.3)
22 (10.7)
LD preoperation
72 (75.0)
24 (21.1)
57 (27.8)
CONCLUSIONS
Postoperative CSF leak
3 (3.1)
3 (2.6)
12 (5.9)
Reoperation required
1 (1.0)
2 (1.8)
5 (2.4)
Endoscopic endonasal repair of CSF leaks and encephaloceles has evolved significantly during the past decade. Safety, efficacy, and versatility have improved dramatically and have surpassed traditional open approaches in these respects. The versatility of different endoscopic approaches through the 4 endonasal corridors allows the endoscopic repair of almost all skull-base defects. The use of vascularized pedicled mucosal flaps has evolved to cover these defects as part of multilayered closure strategies. In addition, mounting evidence suggests the superiority of these flaps—improving closure rates and healing times. As the standard of care evolves, so do our expectations. Spontaneous CSF leaks that may once have recurred at a rate that exceeded 30% are now being closed with >95% primary success. New innovations, new strategies, and new 3-dimensional technologies promise to broaden the appeal of endoscopic endonasal surgery even further.
NSF, nasoseptal flap; LD, lumbar drain; CSF, cerebrospinal fluid.
ventriculoperitoneal (VP) shunt, and use of acetazolamide. Although we place a lumbar drain after surgery for 3 days, we do not use VP shunts or acetazolamide unless the first attempt at endonasal closure fails. We place lumbar drains preoperatively on all patients undergoing CSF leak repair and on all patients with skull-base tumors who are likely to develop an intraoperative CSF leak. However, the uniform placement of a lumbar drain on all patients undergoing endoscopic CSF leak repair is controversial. Casiano and Jassir (9) reported on their series of 33 patients undergoing endoscopic CSF leak repair in 1999. In that series, 32 of 33 (97%) leaks were successfully repaired after 1 procedure without the use of a lumbar drain. However, only 6 patients in this series had noniatrogenic (spontaneous or traumatic) leak etiologies. This is similar to another series from the 1990s by Hughes et al. (17). They reported successful primary endoscopic closure of 16 of 17 (95%) CSF leaks also without the use of a lumbar drain. In that series, 9 patients had
iatrogenic leaks, 6 patients had spontaneous leaks, and 2 patients had traumatic leaks. Postoperative leak rates in these series are comparable with our more recent data (2.9% leak rate) using vascularized pedicled flaps and lumbar drains in all patients; however, there is significant variability in the literature. In their large 800-patient series, Kassam et al. (19) reported a 15.9% leak rate after primary closure, but argued that leak rates are directly proportional to the complexity and etiology of the lesion repaired. Successful closure of patients with preoperative spontaneous CSF leak and BIH is notoriously difficult and management of postoperative ICP is especially important. Postoperative treatment with acetazolamide is widely used in these patients. Acetazolamide is a diuretic that has been shown to reduce CSF production by 48% (8). Woodworth et al. (39) described a paradigm for the treatment of these patients that included 500 mg of acetazolamide twice a day postoperatively on all patients with elevated ICP as evidenced by either elevated lumbar
WORLD NEUROSURGERY 82 [6S]: S86-S94, DECEMBER 2014
REFERENCES 1. Aiden K, Guven K, Sencer S: MRI cisternography with gadolinium-containing contrast medium: its role, advantages and limitations in the investigation of rhinorrhoea. Neuroradiology 46:75-80, 2004. 2. Algin O, Hakyemez B, Gokalp G, Ozcan T, Korfali E, Parlak M: The contribution of 3D-CISS and contrast-enhanced MR cisterography in
www.WORLDNEUROSURGERY.org
S93
PEER-REVIEW REPORTS JEFFREY C. BEDROSIAN ET AL.
detecting cerebrospinal fluid leak in patient with rhinorrhoea. Br J Radiol 83:225-232, 2010. 3. Anand VK, Murali RK, Glasgold MR: Surgical decisions in the management of cerebrospinal fluid rhinorrhoea. Rhinology 33:212-218, 1995. 4. Banks CA, Palmer JN, Chiu AG, O’Malley BW Jr, Woodworth BA, Kennedy DW: Endoscopic closure of CSF rhinorrhea: 193 cases over 21 years. Otolaryngol Head Neck Surg 140:826-833, 2009. 5. Becker SS, Duncavage JA, Russell PT: Endoscopic endonasal repair of difficult-to-access cerebrospinal fluid leaks of the frontal sinus. Am J Rhinol Allergy 23:181-184, 2009. 6. Brown SM, Tabaee A, Singh A, Schwartz T, Anand VK: Three-dimensional endoscopic sinus surgery: feasibility and technical aspects. Otolaryngol Head Neck Surg 138:400-402, 2008. 7. Cappabianca P, Cavallo LM, Colao A, de Divitiis E: Surgical complications associated with the endoscopic endonasal transsphenoidal approach for pituitary adenomas. J Neurosurg 97:293-298, 2002. 8. Carrion E, Hertzog JH, Medlock MD, Hauser GJ, Dalton HJ: Use of acetazolamide to decrease cerebrospinal fluid production in chronically ventilated patients with ventriculopleural shunts. Arch Dis Child 84:68-71, 2001. 9. Casiano RR, Jassir D: Endoscopic cerebrospinal fluid rhinorrhea repair: Is a lumbar drain necessary? Otolaryngol Head Neck Surg 121:745-750, 1999.
ENDOSCOPIC CSF AND ENCEPHALOCELE REPAIR
management based on the Mayo clinic experience from 1970 through 1981. Neurosurgery 16:313-321, 1985.
30. Schick B, Ibing R, Brors D, Draf W: Long-term study of endonasal duraplasty and review of the literature. Ann Otol Rhinol Laryngol 110:142-147, 2001.
17. Hughes RGM, Jones NS, Robertson IJA: The endoscopic treatment of cerebrospinal fluid rhinorrhea: the Nottingham experience. J Laryngol Otol 111:125-128, 1997.
31. Schlosser RJ, Bolger WE: Endoscopic management of cerebrospinal fluid rhinorrhea. Otolaryngol Clin North Am 39:523-538, 2006.
18. Jakimovski D, Bonci G, Attia M, Schwartz T: Incidence and significance of intraoperative CSF leak in endoscopic pituitary surgery using intrathecal fluorescein. World Neurosurg 82:e513-e523, 2014. 19. Kassam AB, Prevedello DM, Carrau RL, Snyderman CH, Thomas A, Gardner P, Zanation A, Duz B, Stefko ST, Byers K, Horowitz MB: Endoscopic endonasal skull base surgery: analysis of complications in the authors’ initial 800 patients. J Neurosurg 114:1544-1568, 2011. 20. Kassam AB, Thomas A, Carrau RL, Snyderman CH, Vescan A, Prevedello D, Mintz A, Gardner P: Endoscopic reconstruction of the cranial base using a pedicled nasoseptal flap. Oper Neurosurg 63:544-553, 2008. 21. Kerr JT, Chu FWK, Bayles SW: Cerebrospinal fluid rhinorrhea: diagnosis and management. Otolaryngol Clin N Am 38:597-611, 2005. 22. Leng LZ, Brown S, Anand VK, Schwartz TH: “Gasket-seal” watertight closure in minimal access endoscopic cranial base surgery. Neurosurgery 62:ONE342-ONE343, 2008.
32. Schwartz TH, Fraser JF, Brown S, Tabaee A, Kacker A, Anand VK: Endoscopic cranial base surgery: classification of operative approaches. Neurosurgery 62:991-1005, 2008. 33. Selcuk H, Albayram S, Ozer H, Ulus S, Sanus GZ, Kaynar MY, Kocer N, Islak C: Intathecal gadolinium-enhanced MR cisternography in the evaluation of CSF leakage. Am J Neuroradiol 31:71-75, 2010. 34. Sen C, Triana A: Cranial chordomas: results of radical excision. Neurosurg Focus 10:E3, 2001. 35. Seth R, Rajasekaran K, Benninger MS, Batra PS: The utility of intrathecal fluorescein in cerebrospinal fluid leak repair. Otolaryngol Head Neck Surg 143:626-632, 2010. 36. Tabaee A, Anand VK, Esposito F, Schwartz ment of spontaneous the lateral sphenoid 1070-1077, 2010.
Cappabianca P, Stamm A, TH: Endoscopic managemeningoencephalocele of sinus. J Neurosurg 112:
37. Tabaee A, Anand VK, Fraser JF, Brown SM, Singh A, Schwartz TH: Three-dimensional endoscopic pituitary surgery. Neurosurgery 64 (5 Suppl 2):288-293, 2009.
10. Castelnuovo P, Dallan I, Pistochini A, Battaglia P, Locatelli D, Bignami M: Endonasal endoscopic repair of Sternberg’s canal cerebrospinal fluid leaks. Laryngoscope 117:345-349, 2007.
23. Mattox DE, Kennedy DW: Endoscopic management of cerebrospinal fluid leaks and cephaloceles. Laryngoscope 100:857-862, 1990.
38. Tabaee A, Placantonakis DG, Schwartz TH, Anand VK: Intrathecal fluorescein in endoscopic skull base surgery. Otolaryngol Head Neck Surg 137:316-320, 2007.
11. Dehdashti AR, Ganna A, Karabatsou K, Gentili F: Pure endoscopic endonasal approach for pituitary adenomas: early surgical results in 200 patients and comparison with previous microsurgical series. Neurosurgery 62:1006-1017, 2008.
24. McCoul ED, Anand VK, Singh A, Nyquist GG, Schaberg MR, Schwartz TH: Long-term outcomes after endoscopic skull base reconstruction with the nasoseptal flap. World Neurosurg 81:136-143, 2014.
39. Woodworth BA, Prince A, Chiu AG, Cohen NA, Schlosser RJ, Bolger WE, Kennedy DW, Palmer JN: Spontaneous CSF leaks: a paradigm for definitive repair and management of intracranial hypertension. Otolaryngol Head Neck Surg 138:715-720, 2008.
25. Nyquist GG, Anand VK, Mehra S, Kacker A, Schwartz TH: Endoscopic endonasal repair of anterior skull base non-traumatic cerebrospinal fluid leaks, meningoceles and encephaloceles. J Neurosurg 113:961-966, 2010.
40. Yang Z, Wang B, Wang C, Liu P: Primary spontaneous cerebrospinal fluid rhinorrhea: a symptom of idiopathic intracranial hypertension? J Neurosurg 115:165-170, 2011.
12. Feiz-Erfan I, Han PP, Spetzler RF, Horn EM, Klopfenstein JD, Porter RW, Ferreira MA, Beals SP, Lettieri SC, Joganic EF: The radical transbasal approach for resection of anterior and midline skull base lesions. J Neurosurg 103:485-490, 2005. 13. Fraser JF, Allen B, Anand VK, Schwartz TH: Threedimensional neurostereoendoscopy: subjective and objective comparison to 2D. Minim Invasive Neurosurg 52:25-31, 2009. 14. Gassner HG, Ponikau JU, Sherris DA, Kern EB: CSF rhinorrhea: 95 consecutive surgical cases with long term follow-up at the Mayo Clinic. Am J Rhinol 13:439-447, 1999.
26. Nyquist GG, Anand VK, Singh A, Schwartz TH: Janus flap: bilateral nasoseptal flaps for anterior skull base reconstruction. Otolaryngol Head Neck Surg 142:327-331, 2010. 27. Ommaya AK, di Chiro G, Baldwin M, Pennybacker JB: Non-traumatic cerebrospinal fluid rhinorrhea. J Neurol Neurosurg Psychiatry 31: 214-225, 1968.
15. Hadad G, Bassagasteguy L, Carrau RL, Mataza JC, Kassam A, Snyderman CH, Mintz A: A novel reconstructive technique after endoscopic expanded endonasal approaches: vascular pedicle nasoseptal flap. Laryngoscope 116:1882-1886, 2006.
28. Placantonakis DG, Tabaee A, Anand VK, Schwartz TH: Safety of low-dose intrathecal fluorescein in endoscopic cranial base surgery. Oper Neurosurg 61:161-166, 2007.
16. Hubbard JL, McDonald TJ, Pearson BW, Laws ER Jr: Spontaneous cerebrospinal fluid rhinorrhea: evolving concepts in diagnosis and surgical
29. Sautter NB, Batra PS, Citardi MJ: Endoscopic management of sphenoid sinus cerebrospinal fluid leaks. Ann Otol Rhinol Laryngol 117:32-39, 2008.
S94
www.SCIENCEDIRECT.com
41. Zuckerman JD, DelGaudio JM: Utility of pre-operative high-resolution CT and intraoperative image guidance in identification of cerebrospinal fluid leaks for endoscopic repair. Am J Rhinol 22:151-154, 2008.
Conflict of interest statement: The authors declare that the article content was composed in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Received 18 September 2013; accepted 24 July 2014 Citation: World Neurosurg. (2014) 82, 6S:S86-S94. http://dx.doi.org/10.1016/j.wneu.2014.07.018 Journal homepage: www.WORLDNEUROSURGERY.org Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2014 Elsevier Inc. All rights reserved.
WORLD NEUROSURGERY, http://dx.doi.org/10.1016/j.wneu.2014.07.018
