Surgical repair of encephaloceles and CSF rhinorrhea

Transcription

1 See the corresponding article in this issue, pp J Neurosurg 113: , 2010 Endoscopic endonasal repair of anterior skull base non-traumatic cerebrospinal fluid leaks, meningoceles, and encephaloceles Clinical article Gu r s t o n G. Ny q u i s t, M.D., 1 Vi j ay K. An a n d, M.D., 1 Sa r a l Me h r a, M.D., M.B.A., 1 Ash u to sh Ka c k e r, M.D., 1 a n d Th e o d o r e H. Sc h w a r t z, M.D. 1,2 Departments of 1 Otorhinolaryngology and 2 Neurological Surgery, Weill Cornell Medical College, New York Presbyterian Hospital, New York, New York Object. The endoscopic endonasal approach has become the preferred technique for CSF leak and encephalocele repair of the anterior skull base. The purpose of this study is to identify patient characteristics; review adjunctive perioperative treatments, reconstruction techniques, and outcomes; and identify risk factors for failure in patients undergoing endoscopic endonasal repair of anterior skull base CSF leaks and encephaloceles. Methods. This is a prospective observational study of patients undergoing endoscopic endonasal repair of a CSF leak between October 2004 and May Twenty-eight consecutive patients underwent 32 procedures. Twenty-two of the patients were women, which represents a statistically significant trend toward the female sex (p < 0.05). The average body mass index (33.9) was significant for obesity. The origin of the skull base defect included the cribriform plate (in 9 cases), fovea ethmoidalis (in 7), combined fovea ethmoidalis/cribriform plate (in 2), lateral sphenoid sinus (in 6), sella (in 4), clivus (in 3), and frontal sinus (in 1). Results. The overall endonasal closure rate was 93.8% (30 of 32 procedures). One failure occurred due to overaggressive postoperative debridement, while the other recurred along the posterior wall of the frontal sinus, and endoscopic repair would have occluded the recess. Conclusions. The endoscopic endonasal approach for the treatment of CSF leaks and encephaloceles of the anterior skull base is the preferred method of repair in the vast majority of cases. The authors 93.8% closure rate in a variety of anatomical locations compares favorably with the transcranial approach and echoes the results of other endoscopic series. (DOI: / JNS08986) Ke y Wo r d s endoscopic approach encephalocele cerebrospinal fluid leak endoscopic skull base Abbreviations used in this paper: BMI = body mass index; VP = ventriculoperitoneal. Surgical repair of encephaloceles and CSF rhinorrhea is recommended to prevent meningitis, intracranial abscess, and pneumocephalus. Traditionally, CSF leaks have been managed via a craniotomy with a 70 80% successful closure rate. 7 Advantages of the transcranial approach include direct visualization of the dural defect, the ability to address associated brain injury, and the potential to use a large vascularized pericranial flap. However, many studies have reported a 40% recurrence rate with this approach and significant patient morbidity including anosmia, frontal lobe retraction, seizures, memory deficits, and intracranial hemorrhage. 1,13,14 In attempt to avoid these complications and improve closure rates, the endonasal endoscopic approach has evolved to address CSF leaks of the anterior skull base. This involves pathology along the cribriform plate (Fig. 1), fovea ethmoidalis, sphenoid bone, or temporal bone. The endoscope not only provides excellent visualization, but outcome studies have demonstrated decreased morbidity and improved closure rates. 3,6,10,13,18 The approach also allows for close postoperative surveillance of the wound site and the ability to identify recurrences. Nevertheless, many neurosurgeons still continue to close CSF leaks of the anterior skull base via a craniotomy, having not been trained in endoscopic endonasal surgery. Cerebrospinal fluid leaks can be divided into posttraumatic, iatrogenic, and spontaneous, the first of which generally resolve without surgical intervention. The lat 961

2 G. G. Nyquist et al. ter 2 categories can be particularly challenging to repair due to their high volume of flow, the difficulty in iden tifying the precise location of the leak, the potential for unrecognized mild increases in intracranial pressure, or congenital thinning of the skull base. The purpose of this study was to review a series of patients who underwent purely endonasal, endoscopic repair of CSF leaks caused by a variety of nontraumatic factors. We define patient characteristics; review adjunctive perioperative treat ments, reconstruction techniques, outcomes; and identify risk factors for failure in patients undergoing endoscopic endonasal repair of anterior skull base CSF leaks and en cephaloceles as an incentive for more neurosurgeons to adopt these techniques in their practice. Methods This is a prospective, observational study of all pa tients undergoing endoscopic repair of a CSF leak caused by a spontaneous meningocele, meningoencephalocele, prior craniotomy, sinus surgery, or remote endoscopic skull base surgery between October 2004 and May Patients with posttraumatic CSF leaks were excluded. Pa tients with CSF leaks caused by recent (< 8 months) endo nasal endoscopic skull base surgery were also excluded. All patients were treated by the senior authors at the In stitute for Minimally Invasive Skull Base and Pituitary Surgery at Weill Cornell Medical College, New York Presbyterian Hospital (V.K.A., A.K., and T.H.S.). The di agnostic methods used to confirm the presence and site of the lesion included any combination of β-2 transferrin testing, rigid nasal endoscopy, CT imaging, MR imaging, cisternography, and intraoperative intrathecal fluorescein. This review evaluated demographic data, location, cause, approach, management, recurrence, and length of follow-up. A total of 32 consecutive endonasal endoscopic sur geries for CSF leak and encephalocele repairs were per formed in 28 patients. The group included 22 women who represent 78.5% of the patients. This represents a statisti cally significant trend toward female sex (p < 0.05, Fisher exact test). The average patient age was 53.6 years (range years). Twenty-four cases (75%) were performed in patients who were overweight or obese (BMI > 25), and 19 (59.4%) were obese (BMI > 30). The average BMI was 33.9 (range ). Skull base defects were repaired through an endo scopic endonasal approach that included a single corridor or any combination of the transsphenoidal, transethmoi dal, transcribriform, transclival, or transpterygoidal corri dors. The details of these approaches have been described in previous articles.16 Intrathecal fluorescein was used to help localize the CSF leak, determine the extent of the defect, and ensure a watertight closure (Fig. 2).19 Skull base reconstruction followed an algorithm based on the size, shape, and location of the defect.17 The first step of the surgery was adequate exposure of the entire defect. The meningocele or encephalocele was then resected, and the bony edges of the defect were defined circumferen tially. This maneuver often worsened the CSF leak but was critical to obtain an eventual watertight seal. Since 962 Fig. 1. Magnetic resonance images demonstrating that meningoceles through the ethmoid roof can project either medially (cribriform plate) (A) or laterally (fovea ethmoidalis) (B) to the vertical attachment of the middle turbinate. Endoscopic views of each of the meningoceles shown in panels A and B demonstrate the lesion medial (C) or lateral (D) to the middle turbinate. the brain is often pushing downward on the defect, there is generally no room for an inlay of fat graft. For small defects, we generally started with an inlay of DuraGuard (Biovascular Corp.) followed by an onlay of fat and fas cia lata that was secured in place with DuraSeal (Conflu ent Surgical). As the defects became larger, our closure technique was more involved. In some circumstances we placed a rigid inlay of Medpor (Porex Surgical), vomer, or even a titanium plate to keep the intracranial contents from herniating through the defect. Use of a vascularized turbinate, nasoseptal, or mucosal graft as a final layer be fore the DuraSeal was often helpful. A gasket seal closure was favored when defects were > 1 cm in diameter.9 This consists of a soft-tissue graft (generally fascia lata) cen tered over the defect with the edges of the graft exceeding the bony defect circumferentially. Next, a piece of rigid material was fashioned to be roughly the size of the bony defect. The rigid graft was centered over the soft-tissue graft so that there was excess soft-tissue graft circumfer entially around the rigid graft. The rigid graft was gently countersunk into the bony defect, forming a watertight gasket seal around the rigid graft. Such a large defect was generally covered with a vascularized flap followed by a final layer of DuraSeal to keep the flap in place. A lumbar drain was placed perioperatively in patients in whom the procedure was deemed a high risk to fail. No formal algorithm was used to prospectively identify these patients, but those who had undergone a previous encephalocele or CSF fistula repair, had a BMI > 30, a large defect, or had an empty sella syndrome identified radiographically tended to have a lumbar drain placed at the time of repair. Lumbar drains, when placed, were opened at a low rate of approximately 5 ml/hour and were generally removed at the end of postoperative Day 1 to facilitate early mobilization of the patient on postopera tive Day 2.

3 Endoscopic encephalocele repair Fig. 2. Intrathecal fluorescein is injected to identify the site of CSF leak. A: Fluorescein-tinted CSF with normal white light. B: Excitation of the fluorescein with blue light leads to emission in the green wavelength that highlights the site of the CSF leak. C: A blocking filter can be placed to remove the blue wavelengths, further highlighting the green wavelengths of the fluorescein. Results Nine surgeries were performed for a CSF leak due to iatrogenic surgical trauma, 2 of which were caused by transcranial skull base surgery and 7 of which were caused by remote endonasal sinus or pituitary surgery. The remaining 23 surgeries were to repair either a spontaneous meningocele or a meningoencephalocele. All patients experienced CSF rhinorrhea as a presenting symptom, 7 (25%) had a history of meningitis, 3 (10.7%) had significant headaches, and 1 patient (3.6%) had a nasal obstruction. Five patients were identified to have an empty sella syndrome on preoperative imaging. The origin of the CSF leaks is shown in Fig. 3. Multiple approaches were used in combination to clearly identify and patch the skull base defect. The approach varied depending on the individual s anatomy and previous surgical interventions, along with the size and location of the defect (Fig. 4). In many cases, multiple approaches were used. The average length of follow-up was 90.5 weeks (range weeks). A total of 24 cases (75%) had a lumbar drain placed perioperatively. Two patients had recurrent CSF leaks at the site of the initial endonasal closure for a success rate of 93.8%. One recurrence was in a patient with a BMI of 31.4 who underwent successful endoscopic closure of her encephalocele until a leak developed immediately after overaggressive debridement by a resident postoperatively. She continues to have a CSF leak and has deferred surgical intervention at this time despite medical counsel otherwise. The second patient was a 71-year-old woman (BMI 49.2) with a large meningoencephalocele (1.5 cm) who underwent successful endonasal repair but then developed a new defect in the posterior table of the frontal sinus, adjacent to our original closure, after a car accident. An attempt was made to repair the defect, but it was thought that an aggressive repair would have a high risk of obstructing the natural outflow from the frontal sinus and result in sinusitis and/or mucocele formation (Fig. 5). After recurrence, the patient was successfully treated via a transcranial approach with a pericranial flap to seal the defect. Two other patients developed recurrent spontaneous CSF leaks in new locations, indicating that the original repair was successful, but the leaks were predisposed to meningocele formation. One of the patients was a 63-year-old woman who underwent resection of a rightsided vestibular schwannoma via a craniotomy. This procedure was complicated by an encephalocele formation through the temporal bone that was successfully repaired through a craniotomy. Four months later, she developed an unrelated encephalocele of the right lateral sphenoid sinus that was successfully repaired endonasally. Two years later, she again developed an encephalocele from the left fovea ethmoidalis that was successfully repaired endonasally as well. On multiple occasions, the patient refused placement of a VP shunt to prevent future leaks. This patient was morbidly obese (BMI 37) and had diabetes in addition to a history of chemotherapy for recurrent breast cancer. The second patient was a 51-year-old woman (BMI 23) with a history of a pituitary macroadenoma that was partially resected and treated with radiation therapy. She had undergone 4 previous CSF leak repairs at an outside institution. Her first endonasal, endoscopic repair at our institution was successful. Six months later the leak recurred at a new site, which was again repaired endonasally, and a VP shunt was placed the following day to prevent further leaks. A year and a half later, she continues to be without a CSF leak. Fig. 3. Pie chart showing the anatomical location of CSF leaks. 963

4 G. G. Nyquist et al. Fig. 4. Pie chart showing the corridors and approaches used to access the CSF leaks. There were 3 cases of postoperative sinusitis that were treated successfully with a course of antibiotics. One patient developed chronic headaches and another a urinary tract infection. A multivariate regression analysis did not identify any significant risk factors for recurrence. Discussion The management of nontraumatic spontaneous or iatrogenic CSF leaks, meningoceles, and meningoencephaloceles has an extensive history characterized by a cornucopia of surgical approaches and different graft materials used to repair the skull base defect. 1,3,6,7,10,11,13,14,18 This study confirms the ability of the endoscopic endonasal approach to reliably identify and repair these cranial defects in multiple locations along the anterior skull base. Our 93.8% closure rate compares favorably with outcomes reported using a craniotomy and echoes the results of other endoscopic series. 1,3,6,7,10,11,13,14,18 These data further solidify the endoscopic approach as the preferred method to address CSF leaks of the anterior skull base that are not associated with additional pathology that requires a transcranial approach, thereby eliminating the risks of brain retraction, seizure, brain hemorrhage, and contusion. The choice of graft material and how to place the graft varies greatly between surgeons. We prefer an inlay technique with autogenous materials (fascia lata) or DuraGuard along with a rigid buttress (vomer or Medpor) and a vascularized flap (nasoseptal or middle turbinate) if the defect is > 1 cm. We have found a final layer of DuraSeal to be helpful in all cases. However, experienced surgeons have yielded similar results using different graft materials and with onlay repair techniques. 2,20 Thus, the choice of graft materials and technique of repair mostly depends on the surgeon s experience and preference. Nonetheless, a few critical practices uniformly improve closure rates. First, the entire defect must be circumferentially identified and accessible for instrumentation. This requires relatively wide exposure. For example, in cases of lateral sphenoid lesions, a transpterygoidal approach is usually required in addition to the transsphenoidal, transethmoidal corridor. The herniated dura should be resected or reduced into the intracranial cavity. Brain that has herniated into the nasal cavity is rarely functional and considered a potential intracranial source of infection Fig. 5. Images obtained in a patient in whom a craniotomy was ultimately required to repair a CSF leak. The patient had a large meningoencephalocele descending through the anterior part of the fovea ethmoidalis (A) filling the entire middle turbinate (B). Endoscopic repair stopped her CSF leak for 9 months, after which she had a car accident and another leak developed. The defect was located at the posterior wall of the frontal sinus (arrows in C and D). Since the only way to access this area from below was through the frontal recess, we did not think we could keep the recess patent and still occlude the defect. Therefore, we elected to perform a craniotomy. if not resected. 5 Another important step is to remove the mucosa surrounding the defect to allow the graft to adhere firmly to the skull base. Our analysis revealed that female sex (p < 0.05) was a statistically significant risk factor for developing an encephalocele or CSF leak. In addition, the average BMI of our patients (33.9) was in the range of obese and significantly higher than a healthy range (BMI between 18.5 and 25). The small number of surgical failures prevented multivariate analysis of our data. However, 2 important teaching points were identified to help prevent recurrence. Conservative nasal debridement in the postoperative period is mandatory. Second, while the endonasal endoscopic approach is the preferred method in the vast majority of cases, one anatomical location that requires careful scrutiny regarding the favored approach is the frontal sinus. This may be a precarious location to reconstruct via an endonasal approach when lesions are far lateral, the frontal recess is narrow, or the reconstruction materials obstruct the outflow from the frontal sinus. Obstruction of the frontal recess due to scarring or from the graft materials often results in sinusitis or mucocele formation, both being technically challenging problems to correct. A transcranial approach with a pericranial flap proved useful in the patient in our series with a posterior table frontal sinus encephalocele. The surgeon must 964

5 Endoscopic encephalocele repair carefully examine the preoperative images to define the extent and location of the skull base defect along with the native anatomy of the frontal sinus to determine the best approach. An area of controversy regarding management involves the use of CSF diversion techniques such as lumbar drainage or a VP shunt. This debate highlights the challenge of treating patients with unrecognized increased CSF pressure. Some authors hypothesize that regardless of the reconstruction technique, patients with increased CSF pressure are at increased risk of persistent or recurrent CSF leak at the reconstruction site or elsewhere along the skull base. While some groups do not favor the use of perioperative lumbar drainage because closure rates may not improve and fear of eliciting pneumocephalus or brain sag, 4,8,12 others use lumbar drainage to measure intracranial pressure to select patients for permanent CSF diversion. 3 Multiple studies have demonstrated higher failure rates in patients with documented hydrocephalus based on CSF opening pressures or presumed hydrocephalus due to conditions such as obesity in patients with a spontaneous CSF leak or an empty sella syndrome. 3,4,6,13,15,18,22 Thus, in our study, we liberally used lumbar drainage at a low drainage rate and short duration without complication to help prevent pooling at the reconstruction site. We do not measure CSF pressure or prophylactically place VP shunts but rather prefer to observe the endonasal closure and reserve shunt placement for patients with recurrent CSF leaks. We calculate the morbidity of shunt placement along with the potential shunt failure rate, which can be as high as 27%, 21 as greater than the possible benefit of preventing a recurrent CSF leak in a small patient population. We aggressively promote weight loss programs in patients with an elevated BMI, prescribe stool softeners, and in rare cases have used acetazolamide to decrease the production of CSF in the perioperative period. Moreover, the addition of vascularized flaps like the nasoseptal or middle turbinate flap to the reconstruction armamentarium may decrease the incidence of recurrence at the reconstruction site and the need for permanent shunting. We recognize that this study is limited by the inherent problems of a retrospective analysis and the use of historical comparisons as a control group. Moreover, the duration of follow-up is limited and may not accurately predict the rate of meningocele or encephalocele recurrence in a disease process that may be chronic. Further prospective multiinstitutional studies are warranted to better understand long-term outcomes, identify patients at risk to experience a recurrence, and the indications for VP shunt treatment in the management of encephaloceles of the anterior skull base. Conclusions The endoscopic endonasal approach for treatment of CSF leaks and encephaloceles of the anterior skull base is the preferred method of repair in the vast majority of cases except the posterior wall of the frontal sinus, where repair may obstruct the frontal recess. Our 93.8% closure rate in a variety of anatomical locations permits the avoidance of a craniotomy and VP shunt placement in the vast majority of cases. These latter treatment modalities should be reserved for only recurrent and refractory cases. Disclaimer The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper. References 1. Aarabi B, Leibrock LG: Neurosurgical approaches to cerebrospinal fluid rhinorrhea. Ear Nose Throat J 71: , Amedee RG, Mann WJ, Gilsbach J: Microscopic endonasal surgery for repair of CSF leaks. Am J Rhinol 7:1 4, Carrau RL, Snyderman CH, Kassam AB: The management of cerebrospinal fluid leaks in patients at risk for high-pressure hydrocephalus. Laryngoscope 115: , Casiano RR, Jassir D: Endoscopic cerebrospinal fluid rhinorrhea repair: is a lumbar drain necessary? Otolaryngol Head Neck Surg 121: , Celin S: Contemporary diagnosis and management of anterior skull base cephalocele and cerebrospinal fluid leaks, in Arriaga MA, Day JD (eds): Neurosurgical Issues in Otolaryngology: Principles and Practice of Collaboration. Philadelphia: Lippincott Williams & Wilkins, 1999, pp Hegazy HM, Carrau RL, Snyderman CH, Kassam A, Zweig J: Transnasal endoscopic repair of cerebrospinal fluid rhinorrhea: a meta-analysis. Laryngoscope 110: , Hubbard JL, McDonald TJ, Pearson BW, Laws ER Jr: Spontaneous cerebrospinal fluid rhinorrhea: evolving concepts in diagnosis and surgical management based on the Mayo Clinic experience from 1970 through Neurosurgery 16: , Kamat AA, Bhattacharyya D, Carroll TA: Brain sag as a cause of postoperative neurological deterioration following anterior cranial fossa floor repair for post traumatic cerebrospinal fluid rhinorrhoea. Br J Neurosurg 21: , Leng LZ, Brown S, Anand VK, Schwartz TH: Gasket-seal watertight closure in minimal-access endoscopic cranial base surgery. Neurosurgery 62 (5 Suppl 2):ONSE342 ONSE343, Lindstrom DR, Toohill RJ, Loerhl TA, Smith TL: Management of cerebrospinal fluid rhinorrhea: the Medical College of Wisconsin experience. Laryngoscope 114: , Locatelli D, Rampa F, Acchiardi I, Bignami M, De Bernardi F, Castelnuovo P: Endoscopic endonasal approaches for repair of cerebrospinal fluid leaks: nine-year experience. Neurosurgery 58 (4 Suppl 2):ONS-246 ONS-257, Mirza S, Saeed SR, Ramsden RT: Extensive tension pneumocephalus complicating continuous lumbar CSF drainage for the management of CSF rhinorrhoea. ORL J Otorhinolaryngol Relat Spec 65: , Park JI, Strelzow VV, Friedman WH: Current management of cerebrospinal fluid rhinorrhea. Laryngoscope 93: , Ray BS, Bergland RM: Cerebrospinal fluid fistula: clinical aspects, techniques of localization, and methods of closure. J Neurosurg 30: , Schlosser RJ, Bolger WE: Significance of empty sella in cerebrospinal fluid leaks. Otolaryngol Head Neck Surg 128:32 38, Schwartz TH, Fraser JF, Brown S, Tabaee A, Kacker A, Anand VK: Endoscopic cranial base surgery: classification of operative approaches. Neurosurgery 62: , Tabaee A, Anand VK, Brown SM, Lin JW, Schwartz TH: 965

6 G. G. Nyquist et al. Algorithm for reconstruction after endoscopic pituitary and skull base surgery. Laryngoscope 117: , Tabaee A, Kassenoff TL, Kacker A, Anand VK: The efficacy of computer assisted surgery in the endoscopic management of cerebrospinal fluid rhinorrhea. Otolaryngol Head Neck Surg 133: , Tabaee A, Placantonakis DG, Schwartz TH, Anand VK: Intrathecal fluorescein in endoscopic skull base surgery. Otolaryngol Head Neck Surg 137: , Weber R, Keerl R, Draf W, Schick B, Mosler P, Saha A: Management of dural lesions occurring during endonasal sinus surgery. Arch Otolaryngol Head Neck Surg 122: , Wu Y, Green NL, Wrensch MR, Zhao S, Gupta N: Ventriculoperitoneal shunt complications in California: 1990 to Neurosurgery 61: , Zweig JL, Carrau RL, Celin SE, Schaitkin BM, Pollice PA, Snyderman CH, et al: Endoscopic repair of acquired encephaloceles, meningoceles, and meningo-encephaloceles: predictors of success. Otolaryngol Head Neck Surg 123: , 2000 Manuscript submitted August 5, Accepted October 5, Please include this information when citing this paper: published online November 20, 2009; DOI: / JNS Address correspondence to: Theodore H. Schwartz, M.D., De partment of Neurological Surgery, Weill Cornell Medical College, New York Presbyterian Hospital, 525 East 68th Street, Box #99, New York, New York