The Incidence of Cerebrospinal Fluid Leak after Vestibular Schwannoma Surgery

Transcription

1 Otology & Neurotology 25: , Otology & Neurotology, Inc. The Incidence of Cerebrospinal Fluid Leak after Vestibular Schwannoma Surgery Samuel H. Selesnick, Jeffrey C. Liu, Albert Jen, and Jason Newman Department of Otorhinolaryngology, Weill College of Medicine of Cornell University, New York, New York, U.S.A. Objective: To review the incidence of cerebrospinal fluid leak after vestibular schwannoma removal reported in the literature. Data Sources: MEDLINE and PubMed literature search using the terms acoustic neuroma or vestibular schwannoma, and cerebrospinal fluid leak or cerebrospinal fluid fistula covering the period from 1985 to the present in the English language literature. A review of bibliographies of these studies was also performed. Selection: Criteria for inclusion in this meta-analysis consisted of the availability of extractable data from studies presenting a defined group of patients who had undergone primary vestibular schwannoma removal and for whom the presence and absence of cerebrospinal fluid leakage was reported. Studies reporting combined approaches were excluded. No duplications of patient populations were included. Twenty-five studies met the inclusion criteria. Data Extraction: Quality of the studies was determined by the design of each study and the ability to combine the data with the results of other studies. All of the studies were biased by their retrospective, nonrandomized nature. Data Synthesis: Significance (p < 0.05) was determined using the 2 test. Conclusions: Cerebrospinal fluid leak occurred in 10.6% of 2,273 retrosigmoid surgeries, 9.5% of 3,118 translabyrinthine surgeries, and 10.6% of 573 middle fossa surgeries. The type of cerebrospinal fluid leak was not associated with surgical approach. Meningitis was significantly associated with cerebrospinal fluid leak (p < 0.05). Age and tumor size were not associated with cerebrospinal fluid leak. Key Words: Acoustic Cerebrospinal fluid Meta-analysis Neuroma Operative Otologic surgical Surgical. Otol Neurotol 25: , Over the past several decades, the expected outcomes from the surgical care of vestibular schwannomas have evolved. Whereas early attempts at surgical management of vestibular schwannomas left patients with significant cranial nerve and neurologic disability, present advances in microsurgical techniques have resulted in minimal postoperative morbidity. Today, both facial nerve function and, under favorable circumstances, hearing can be preserved in patients with small tumors. However, despite these advances, cerebrospinal fluid () leak may still complicate the postoperative course. Cerebrospinal fluid leak can lead to meningitis (1,2), the need for reoperation (3), and increased length of stay in the hospital (3,4). Despite the introduction of many methods intended to prevent leak during translabyrinthine, suboccipital, and middle cranial fossa surgery for vestibular schwannomas, leaks may still occur, and the rate is highly Address correspondence and reprint requests to Samuel H. Selesnick, M.D., F.A.C.S., Department of Otorhinolaryngology, Weill College of Medicine of Cornell University, Starr Building, Suite 541, 520 East 70th Street, New York, NY 10021, U.S.A.; shselen@ mail.med.cornell.edu variable. A review of the current literature reveals many proposed alternatives or adjuncts to the standard methods of surgical closure to further decrease the rate of postoperative leak. These methods include the use of fat, fascia, bone wax (5), tissue adhesives (6), and biomaterials (7 9), and even the use of endoscopes to visualize potential sites for leakage (10). The goal of this study was to define the incidence of leak and to explore prophylactic measures that may be used in its prevention. PATIENTS AND METHODS A meta-analysis was performed. A PubMed and MEDLINE search of all articles published in English since 1985 using the keywords cerebrospinal fluid leak or cerebrospinal fistula, and acoustic neuroma or vestibular schwannoma was performed. The criteria for study inclusion were publication of the details of defined groups of patients who had vestibular schwannoma removal and for when rates of leakage were reported. Only retrosigmoid (suboccipital), translabyrinthine, and middle fossa approaches were considered; no combined approaches were included. Studies that reported other approaches or reported nonvestibular schwannomas in the cerebellopontine angle were included only if data on leaks after vestibular schwannoma excision using only the aforementioned approaches could be extracted. A review of bibliographies of selected studies was also performed. Twenty-five studies met 387

2 388 S. H. SELESNICK ET AL. Approach TABLE 1. leaks by surgical approach studies the inclusion criteria. Studies that differentiated the types of leakage (incisional/wound, rhinorrhea, otic) were further examined. Fourteen met the inclusion criteria. Studies by the same authors were examined to exclude overlapping patient populations. The use of prophylactic techniques to prevent leakage was assessed. Data were segregated by either translabyrinthine or retrosigmoid approach. Finally, studies presenting leakage and meningitis rates were pooled. Studies were chosen independent of the type of surgery (retrosigmoid, translabyrinthine, or middle fossa), the type of meningitis (aseptic versus bacterial), and the type of leak. A 2 analysis was performed on the pooled data. RESULTS leaks Retrosigmoid 14 2, Translabyrinthine 17 3, Middle fossa p 0.31 Cerebrospinal fluid leak was first examined by surgical approach. For the retrosigmoid (suboccipital) approach, 14 studies were pooled examining 2,273 (Tables 1 and 2); 10.6% of these (242 cases) resulted in some sort of leakage. For the translabyrinthine approach, 17 studies were pooled examining 3,118 (Tables 1 and 3). The total leak rate was 9.5% (295 cases). Finally, the middle fossa approach was examined in six studies for a total of 573 ; 10.6% (61 cases) of resulted in leakage (Table 1). Overall, the difference between the data from all three was not significant (p 0.31, 2 test). Next, studies that reported leakage by type were pooled and examined (Table 4). The three major types of leakage reported were rhinorrhea, incisional, and otic. In the retrosigmoid approach, 10 studies were examined for a pooled total of 1,944 ; 11.4% of these resulted in leakage, with 6.1% (range, %) of the total manifesting as rhinorrhea, 4.9% (range, %) as incisional leakage, and 0.4% (range, 0 3.9%) as otic leakage. In the translabyrinthine approach, 10 studies were examined for a total of 1,599. Of these, 10.3% of resulted in leakage, with 6.0% (range, %) resulting in rhinorrhea, 4.8% (range, %) of the total resulting in incisional leakage, and 0.3% (range, 0 0.9%) resulting in otic leakage. Differences in frequency between nasal, wound, and otic leak were not significant (p 0.90, 2 test). Prophylactic techniques intended to prevent leak performed at the time of the surgery varied from study to study. In the retrosigmoid approach, 10 studies discussed prophylactic techniques used intraoperatively (Table 5). All studies reported the use of bone wax, except one study that used hydroxyapatite cement. In closing the surgical defect, three studies reported the use of a free fat graft, two studies reported using temporalis muscle, and one study reported using grafts from both locations. Autologous fibrin glue was also used in two capacities either to aid in air cell closure or as a sealant for a temporalis or fascia lata graft (Table 6). No prophylactic technique correlated exclusively with a lower rate of leakage. In the translabyrinthine approach (Table 7), almost all authors, 14 of 16 (88%), reported the use of a free fat graft to pack the mastoid defect. The other two authors used either hydroxyapatite cement or ionomeric cement for closure. Fibrin glue was also used in 38% of studies (Table 8). In these studies, fibrin glue was used either to seal the autologous fat graft or to hold a muscle/fascia TABLE 2. leak and types for the retrosigmoid approach leaks Rhinorrhea Incisional Becker et al., 2003 (15) a 9 2 Leonetti et al., 2001 (25) Brennan et al., 2001 (11) Gal & Bartels, 1999 (5) 35 1 Magliulo et al., 1998 (17) Samii & Matthies, 1997 (26) 1, Valtonen et al., 1997 (10) Fishman et al., 1996 (27) Nutik & Korol, 1995 (28) Gillman & Parnes, 1995 (13) 10 2 Hoffman, 1994 (16) Bryce et al (3) Kemink et al., 1990 (29) 93 2 Mangham, 1988 (22) Total 2, (10.6) 119 (6.5) 96 (5.2) 8 (0.4) Total number of of studies reporting leak by type of leak: 1,944. a includes one combined rhinorrhea/incisional leak. Otic

3 LEAK AFTER VESTIBULAR SCHWANNOMA SURGERY 389 TABLE 3. leak and types for the translabyrinthine approach leaks Rhinorrhea Incisional Becker et al., 2003 (15) a 7 7 Arriaga & Chen 2002 (7) Brennan et al., 2001 (11) Leonetti et al., 2001 (25) Mass et al., 1999 (30) Gal & Bartels, 1999 (5) 27 0 Magliulo et al., 1998 (17) Fishman et al., 1996 (27) Celikkanat et al., 1995 (12) Gillman & Parnes, 1995 (13) 52 6 Hoffman, 1994 (16) Pulec, 1994 (31) Rodgers & Luxford, 1993 (4) Bryce et al., 1991 (3) Hardy et al., 1989 (2) Tos et al., 1988 (14) Mangham, 1988 (22) Total 3, (9.5) 97 (6.5) 76 (5.1) 5 (0.3) Total number of of studies reporting leak by type of leak: 1,599. a includes one combined rhinorrhea/incisional leak. Otic graft in place, with some studies using both techniques. In one study, fibrin glue was also used to fix a piece of preserved dura during closure. No single technique was associated with a leak percentage less than the overall average of 9.5%. Studies using the middle fossa approach had insufficient data that could be extracted for analysis. The frequency of meningitis in patients with leak were tabulated in the 13 studies that fulfilled the inclusion criteria (Table 9). Overall, 2,316 were pooled, of which 86 were complicated by meningitis. Thirty-four (40%) of these had no leak, whereas 52 (60%) were associated with leak. These data were significant (p < , 2 test). Of the 246 leaks in these studies, 34 (14%) developed meningitis. Eight of the 13 studies used a lumbar drain during leak management, one study did not, and three studies had no data regarding the use of lumbar drain or reported no leaks. Of all the studies examined, only seven examined the impact of tumor size on rates of leak. Brennan et al. (11) reported a significant (p 0.001) association of leak with tumor size in the retrosigmoid approach but not in the translabyrinthine approach. Bryce et al. (3) reported a significant (p 0.018) association between leak and tumor size. However, these authors included combined surgical approaches, such as the translabyrinthine/middle fossa approach, in their analysis, which they noted to have a higher leak rate overall. Five studies (10,12 15) reported no significant association between tumor size and leak, so that there is little agreement regarding the ability of tumor size to predict the development of a postoperative leak. Of all studies examined, only four examined the relationship between age and leak (4,15 17). All studies showed no significant relationship (Figs. 1 3). DISCUSSION In this meta-analysis, we have pooled the patient information from 25 studies on leak after vestibular schwannoma surgery. Only seven studies explored the relationship between tumor size and rates of leak. Two of these studies (3,11) showed a significant association. Because individual tumor sizes in each study are not available, we cannot pool the data to examine for significance across studies. All studies examining the relationship between age and leak were insignifi- TABLE 4. Types of leak by surgical approach Approach studies leaks Rhinorrhea Incisional Otic Retrosigmoid 10 1, Range Translabyrinthine 10 1, Range p 0.90

4 390 S. H. SELESNICK ET AL. TABLE 5. Reported prophylactic surgical techniques: retrosigmoid a leaks % Bone wax Free fat graft Temporalis muscle pack Lumbar drainage Fascia lata or temporalis fascia Postoperative mastoid pressure dressing Hydroxyapatite cement Kamerer et al., 1994 (18) 1 2? X Gal & Bartels, 1999 (5) X Brennan et al., 2001 (11) X X Samii & Matthies, 1997 (26) 1, X X Becker et al., 2003 (15) X X X Bryce et al., 1991 (3) X X Nutik & Korol, 1995 (28) X X Valtonen et al., 1997 (10) X X Hoffman, 1994 (16) X X Fishman et al., 1996 (27) X X X a Xs denote studies reporting use of a technique. Absence does not necessarily indicate technique not performed. Temporalis muscle packing: all muscle was placed in middle ear: aditus, mastoid, attic, epitympanum. Only in one case (Becker et al., 2003) was it placed in the internal acoustic canal. Not shown: endoscopic air cell mapping (Valtonen et al., 1997; Becker et al., 2003). cant (4,15 17). There were no preoperative indicators that would predict the development of a postoperative leak. The difference in leak rate between the three surgical approaches examined (translabyrinthine, retrosigmoid, middle fossa) was not significant (p 0.31), so that the surgical approach to vestibular schwannoma does not influence the likelihood of cerebrospinal leakage. Ten studies reported rates of leakage by type (rhinorrhea, incisional, otic) via each surgical approach. Because so much more of the pneumatized temporal bone is opened in the translabyrinthine approach, one would assume that there would be a higher rate of rhinorrhea. However, analysis showed that when a leak occurred, there was no significant difference in the frequency of the types of leak when comparing translabyrinthine and retrosigmoid approaches (p 0.90). Rhinorrhea, incisional, and otic leak make up the same percentage of leaks in both approaches. There were inadequate data to study the middle fossa approach. The prophylactic techniques for prevention of leak are varied. In the retrosigmoid approach, numerous surgical techniques have been used, such as endoscopic mapping of exposed air cells, free fat packing, temporalis muscle packing, fibrin glue, hydroxyapatite cement obliteration, and bone wax obliteration. Unfortunately these techniques were analyzed in just a few articles, so the data were insufficient to suggest that a single technique was associated with a lower incidence of leakage. Overall, the most commonly used prophylactic techniques were obliteration of open mastoid and petrous apex air cells by bone wax and adequate packing of the surgical defect, either with a free fat graft or with temporalis muscle. In the translabyrinthine approach, techniques such as eustachian tube obliteration with fascia lata, temporalis fascia packing, or temporalis muscle packing, although popular, were not associated with studies reporting leaks less than the pooled rate found in this study of 9.5%. Approximately one-third of studies regarding the translabyrinthine approach and the retrosigmoid approach reported the use of autologous fibrin glue in the closure of the primary operation. Specifically, eight studies from 1994 to the present reported leak rates of 6.2 to 21% (Tables 6 and 8). One large study has reported the use of fibrin glue and a historical control group. This study of 492 patients by Lebowitz et al. (6) examined the use of fibrin glue in vestibular schwannoma surgery. In both the retrosigmoid and translabyrinthine approaches, the rate of leak with fibrin glue was shown to be no different than the rate without the use of fibrin glue. Thus, although fibrin glue was prophylactically used by many authors in our study series, this large study by Lebowitz et al. (6) suggests that the use of fibrin glue in addition to routine prophylactic measures is not effective at reducing the incidence of leak after vestibular schwannoma surgery. This study by Lebowitz et al. was not included in the pooled meta-analysis because the TABLE 6. Studies using prophylactic fibrin glue in surgery: retrosigmoid Type leak % Temporalis or fascia lata graft sealant Samii & Matthies, 1997 (26) RS 1, X Hoffman, 1994 (16) RS X Fishman et al., 1996 (27) RS X Lebowitz et al. (6) were not included in this table because the patient population overlaps with Hoffman (16)., cerebrospinal fluid; RS, retrosigmoid. Closure of air cells

5 LEAK AFTER VESTIBULAR SCHWANNOMA SURGERY 391 TABLE 7. Reported prophylactic surgical techniques: translabyrinthine a leaks % Free fat graft Temporalis muscle pack Fascia lata or temporalis fascia ET packing Postoperative mastoid pressure dressing Bone wax ET obliteration Hydroxyapatite cement Gal & Bartels 1999 (5) X X Kamerer et al., 1994 (18) X Helms & Geyer, 1994 (9) X Pulex, 1994 (31) X X X X X X Celikkanat et al., 1995 (12) X X X Rodgers & Luxford, 1993 (4) X X Mass et al., 1999 (30) X X Arriaga & Chen, 2002 (7) X X Brennan et al., 2001 (11) X X X Tos et al., 1988 (14) X X Bryce et al., 1991 (3) X X X X X Gillman & Parnes, 1995 (13) X X X X Hardy et al., 1989 (2) X X X Becker et al., 2003 (15) X Nutik & Korol, 1995 (28) X X Hoffman, 1994 (16) X X X a Xs denote studies reporting use of a technique. Absence does not necessarily indicate technique not performed. All temporalis muscle packing refers to packing at one or more of the following sites: aditus, mastoid, attic, epitympanum. All eustachian tube packing was performed with temporalis muscle, except Hardy et al., who used fat. Not shown: ionomeric cement (Helms et al., 1994)., cerebrospinal fluid; ET, eustachian tube. patient population overlapped with the reported Hoffman study (16). Although no specific technique used by multiple studies was associated with a low rate of leakage, a few single studies using new techniques did report significantly lower leakage rates. Kamerer and colleagues (18) used hydroxyapatite cement for closure in seven vestibular schwannoma cases. Four translabyrinthine, two middle fossa, and one retrosigmoid approaches were performed. In all these cases, when hydroxyapatite cement was used in place of bone wax to fill exposed air cells adjacent to the internal auditory canal and tegmen, no cases of leak were reported. In another study reported by Arriaga and Chen (7), 54 patients with abdominal fat graft reconstruction were compared with 54 patients who underwent hydroxyapatite cement reconstruction of translabyrinthine craniotomies for vestibular schwannoma. Although the abdominal fat graft group had 7 leaks out of 54 (12.9%), the hydroxyapatite group only had 2 leaks (3.7%). These studies suggest that the use of hydroxyapatite cement may be a promising adjunct in reducing the rates of leakage in translabyrinthine acoustic surgery. A study by Valtonen and colleagues (10) reported the use of endoscope-assisted waxing of the petrous apex air cells in patients undergoing a suboccipital craniotomy for vestibular schwannoma surgery. In this technique, before closure, the surgeon used a 30-degree rigid endoscope to visualize exposed air cells in the petrous apex before sealing them. Twenty-four patients underwent surgery using the endoscopic technique and were compared with a control group of 38 patients who received the same surgery without endoscopic petrous apex visualization. The control group had a leak rate of 18.4%, whereas the experimental group showed a 4.2% leak rate. Helms and Geyer (9) used ionomeric cement to aid in closure of patients after translabyrinthine vestibular schwannoma removal. In this series, the defect was first filled with temporalis muscle impregnated with fibrin glue. Then, the petrous apex was sealed with preserved dura. Finally, an approximately 2- to 3-mm-thick layer of ionomeric cement was applied on top of the dura. In their series of five patients, none developed a leak. It is important to note that two studies have demonstrated that ionomeric cement is potentially harmful. Lubben and Geyer (19) reported that ionomeric cement was cytotoxic to mouse fibroblasts in cell culture. Fur- TABLE 8. Studies using prophylactic fibrin glue in surgery: translabyrinthine Type leak % Fat graft seal Temporalis or fascia lata graft sealant Preserved dura fixation Helms & Geyer 1994 (9) TL X X Celikkanat et al., 1995 (12) TL X Tos et al., 1988 (14) TL X Gillman & Parnes, 1995 (13) TL X Hardy et al., 1989 (2) TL X X Hoffman, 1994 (16) TL X X Lebowitz et al. (6) was not included in this table because the patient population overlaps with Hoffman (16)., cerebrospinal fluid; TL, translabyrinthine.

6 392 S. H. SELESNICK ET AL. TABLE 9. Studies reporting meningitis and leakage + Menin Menin + +Menin +Menin Arriaga & Chen, 2002 (7) Y Stidham & Roberson, Y Mass et al., 1999 (30) Y Valtonen et al., 1997 (10) Y Slattery et al., 1997 (35) Y Weber & Gantz, 1996 (33) Y Meyerson et al., 1996 (20) ND Gillman & Parnes, 1995 (13) Y Pulec, 1994 (31) ND Hoffman, 1994 (16) Y Rodgers & Luxford, 1993 (4) N Kanzaki et al., 1991 (32) N Mangham, 1988 (22) ND Total 212 2, Total : 2,316. p < ( 2 )., cerebrospinal fluid leak; Menin, meningitis; ND, not discussed. Lumbar drain thermore, a study by Kupperman and Tange (8) examined the postoperative outcomes of 23 patients after the use of ionomeric cement in middle ear surgery. Six of the 23 patients were found to have a severe tissue reaction with middle ear infection and otorrhea; 9 patients were found to have extrusion of the ionomeric cement after 5 years. On the basis of these studies, we recommend that ionomeric cement not be used in the closure of craniotomy defects. Another study by Meyerson et al. (20) reported the use of polytetrafluoroethylene (PTFE) sponges in prevention of leak. In their approach, the authors packed the eustachian tube with PTFE felt before free muscle and fat graft closure of the middle ear. In their series, no cases of rhinorrhea were reported, and only one case of incisional and otic leakage each was reported in 25 patients. The risk of the use of the PTFE sponge is not known. On the basis of our data, the best prophylactic measures in preventing leakage of all types in translabyrinthine vestibular schwannoma surgery are adequate closure of air cells and the middle ear defect. The study by Valtonen et al. using endoscope-assisted closure supports the concept that aggressive closure of air cells prevents tracking and subsequent rhinorrhea. Adequate packing of the surgical defect, most commonly with abdominal fat, creates a tissue seal, often preventing incisional leakage. Because the eustachian tube is the final common pathway for the development of rhinorrhea, we recommend routine eustachian tube obliteration during vestibular schwannoma removal. This can be achieved through a facial recess approach with removal of the incus. In this way, adequate access is obtained while still keeping the external auditory canal wall intact. In some cases, it may also be reasonable to use the transotic approach as described by Jenkins and Fisch (21) as in patients for large vestibular schwannomas. In this approach, the external auditory canal wall is removed and the external auditory meatus is oversewn. A more direct access to the eustachian tube orifice is achieved and direct obliteration can be performed more thoroughly. In patients in whom the surgeon suspects there is a high risk of leak, such as patients with a highly pneumatized mastoid on a preoperative computed tomographic scan of the temporal bone, a more thorough procedure may be indicated such as a transaural transnasal approach (see this issue of Otology and Neurotology: Management Options for Cerebrospinal Fluid Leak after Vestibular Schwannoma Surgery, and Introduction of an Innovative Treatment ). Hoffman (16) noted that the association between leak and meningitis was unclear. He noted that Bryce et al. (3) and Tos and Thomsen (1) all reported a significant association between meningitis and leak. Yet, Rodgers and Luxford (4) and Mangham (22) were unable to find a significant association. In our meta-analysis of 13 pooled studies, 14% of leaks were associated with meningitis, significantly higher than the meningitis rate (3%) in patients without leaks (p < ). Eight of 13 studies examined incorporated the use of a lumbar drain during leak management; however, the data could not be extracted to determine the relationship between lumbar drainage and meningitis. Previous studies have reported a meningitis rate between 2% and 4% (23 35) in association with lumbar drain usage. We therefore conclude that leak, and not merely placement of a lumbar drain, contributes to the incidence of meningitis. CONCLUSION Despite the use of purposeful prophylactic methods, cerebrospinal fluid leak continues to be an infrequent but persistent complication after vestibular schwannoma resection. In a meta-analysis of data presented in this article, leak occurs in 10.6% of retrosigmoid, 9.5% of translabyrinthine, and 10.6% of middle fossa surgical. Analysis of prophylactic techniques in the translabyrinthine and retrosigmoid approaches failed to

7 LEAK AFTER VESTIBULAR SCHWANNOMA SURGERY 393 demonstrate a definitive technique for elimination of leak. Finally, meningitis was shown to be strongly associated with leakage. Acknowledgment: The authors thank Jonathan Victor, M.D., for assistance with the statistical analysis. REFERENCES 1. Tos M, Thomsen J. The price of preservation of hearing in acoustic neuroma surgery. Ann Otol Rhinol Laryngol 1982;91(3 Pt 1): Hardy DG, Macfarlane R, Baguley D, et al. Surgery for acoustic neurinoma: an analysis of 100 translabyrinthine operations. J Neurosurg 1989;71: Bryce GE, Nedzelski JM, Rowed DW, et al. Cerebrospinal fluid leaks and meningitis in acoustic neuroma surgery. Otolaryngol Head Neck Surg 1991;104: Rodgers GK, Luxford WM. Factors affecting the development of cerebrospinal fluid leak and meningitis after translabyrinthine acoustic tumor surgery. Laryngoscope 1993;103: Gal TJ, Bartels LJ. Use of bone wax in the prevention of cerebrospinal fluid fistula in acoustic neuroma surgery. Laryngoscope 1999;109: Lebowitz RA, Hoffman RA, Roland JT Jr, et al. Autologous fibrin glue in the prevention of cerebrospinal fluid leak following acoustic neuroma surgery. Am J Otol 1995;16: Arriaga MA, Chen DA. Hydroxyapatite cement cranioplasty in translabyrinthine acoustic neuroma surgery. Otolaryngol Head Neck Surg 2002;126: Kupperman D, Tange RA. Ionomeric cement in the human middle ear cavity: long-term results of 23 cases. Laryngoscope 2001;111: Helms J, Geyer G. Closure of the petrous apex of the temporal bone with ionomeric cement following translabyrinthine removal of an acoustic neuroma. J Laryngol Otol 1994;108: Valtonen HJ, Poe DS, Heilman CB, et al. Endoscopically assisted prevention of cerebrospinal fluid leak in suboccipital acoustic neuroma surgery. Am J Otol 1997;18: Brennan JW, Rowed DW, Nedzelski JM, et al. Cerebrospinal fluid leak after acoustic neuroma surgery: influence of tumor size and surgical approach on incidence and response to treatment. J Neurosurg 2001;94: Celikkanat SM, Saleh E, Khashaba A, et al. Cerebrospinal fluid leak after translabyrinthine acoustic neuroma surgery. Otolaryngol Head Neck Surg 1995;112: Gillman GS, Parnes LS. Acoustic neuroma management: a six-year review. J Otolaryngol 1995;24: Tos M, Thomsen J, Harmsen A. Results of translabyrinthine removal of 300 acoustic neuromas related to tumor size. Acta Otolaryngol Suppl 1988;452: Becker SS, Jackler RK, Pitts LH. Cerebrospinal fluid leak after acoustic neuroma surgery: a comparison of the translabyrinthine, middle fossa, and retrosigmoid approaches. Otol Neurotol 2003; 24: Hoffman RA. Cerebrospinal fluid leak following acoustic neuroma removal. Laryngoscope 1994;104(1 Pt 1): Magliulo G, Sepe C, Varacalli S, et al. Cerebrospinal fluid leak management after cerebellopontine angle surgery. J Otolaryngol 1998;27: Kamerer DB, Hirsch BE, Snyderman CH, et al. Hydroxyapatite cement: a new method for achieving watertight closure in transtemporal surgery. Am J Otol 1994;15: Lubben B, Geyer G. Toxicity of glass ionomer cement. Laryngorhinootologie 2001;80: Meyerson LR, Monsell EM, Rock JP. Preventive management of cerebrospinal fluid leakage in translabyrinthine surgery. Laryngoscope 1996;106(5 Pt 1): Jenkins HA, Fisch U. The transotic approach to resection of difficult acoustic tumors of the cerebellopontine angle. Am J Otol 1980;2: Mangham CA. Complications of translabyrinthine vs suboccipital approach for acoustic tumor surgery. Otolaryngol Head Neck Surg 1988;99: Shapiro SA, Scully T. Closed continuous drainage of cerebrospinal fluid via a lumbar subarachnoid catheter for treatment or prevention of cranial/spinal cerebrospinal fluid fistula. Neurosurgery 1992;30: Coplin WM, Avellino AM, Kim DK, et al. Bacterial meningitis associated with lumbar drains: a retrospective cohort study. J Neurol Neurosurg Psychiatry 1999;67: Leonetti J, Anderson D, Marzo S, et al. Cerebrospinal fluid fistula after transtemporal skull base surgery. Otolaryngol Head Neck Surg 2001;124: Samii M, Matthies C. Management of 1000 vestibular schwannomas (acoustic neuromas): surgical management and results with an emphasis on complications and how to avoid them. Neurosurgery 1997;40: Fishman AJ, Hoffman RA, Roland JT Jr, et al. Cerebrospinal fluid drainage in the management of leak following acoustic neuroma surgery. Laryngoscope 1996;106: Nutik SL, Korol HW. Cerebrospinal fluid leak after acoustic neuroma surgery. Surg Neurol 1995;43: Kemink JL, LaRouere MJ, Kileny PR, et al. Hearing preservation following suboccipital removal of acoustic neuromas. Laryngoscope 1990;100: Mass SC, Wiet RJ, Dinces E. Complications of the translabyrinthine approach for the removal of acoustic neuromas. Arch Otolaryngol Head Neck Surg 1999;125: Pulec JL. Technique to avoid cerebrospinal fluid otorhinorrhea with translabyrinthine removal of acoustic neuroma. Laryngoscope 1994;104(3 Pt 1): Kanzaki J, Ogawa K, Tsuchihashi N, et al. Postoperative complications in acoustic neuroma surgery by the extended middle cranial fossa approach. Acta Otolaryngol Suppl 1991;487: Weber PC, Gantz BJ. Results and complications from acoustic neuroma excision via middle cranial fossa approach. Am J Otol 1996;17: Shelton C, Brackmann DE, House WF, et al. Middle fossa acoustic tumor surgery: results in 106 cases. Laryngoscope 1989;99: Slattery WH III, Brackmann DE, Hitselberger W. Middle fossa approach for hearing preservation with acoustic neuromas. Am J Otol 1997;18: