A Comparison of Management Strategies for Cerebrospinal Fluid Leaks following Vestibular Schwannoma Removal

Introduction: Cerebrospinal fluid (CSF) leak is a frustrating complication following vestibular schwannoma (VS) removal. We analyzed the outcome comparing CSF diversion via lumbar drainage versus direct operative repair in patients who developed CSF leaks after VS surgery.

Methods: A prospectively maintained database of patients undergoing VS resection between 1999–2013 was reviewed. Patients who developed postoperative CSF otorrhea or rhinorrhea were included. Baseline characteristics gathered included age, gender, initial surgical approach, median follow up, time to leak onset and leak duration. Patients were managed with either lumbar drainage or return to the OR for attempt at direct repair of the leak. Patients were further grouped based on the timing of the onset of CSF leak (acute = during initial hospital stay, perioperative = less than 30 days after VS removal, delayed = greater than 30 days). Patients with delayed CSF leaks were excluded from further analysis as they were always managed with reoperation. The primary outcome measure was treatment failure defined as the need for additional intervention (further operative repair or ventriculoperitoneal shunt for CSF diversion). Secondary outcome measures included hospital days until cessation of CSF leak. A qualitative assessment of standard patient costs and actual charges was performed.

Results: Of the 462 patients that underwent VS resection during the study period, 50 (10.8%) developed CSF leaks. Thirteen patients were excluded from further analysis as they developed cutaneous CSF fistulas managed by simply oversewing the wound. An additional 5 patients were excluded as they developed delayed leaks. Thus a total of 32 patients were included for further analysis. Of these 32 perioperative leaks, 20 (63%) were acute. The median time to leak onset was 4 days (range 0–27 days). Median follow up was 26 months (range 1–136 months). The overall recurrence rate was 24% (N = 9). Lumbar drainage was the primary management strategy in 23 patients compared with operative intervention in 10 patients and combined strategies in 4 patients. The overall success rate of lumbar drainage as primary treatment was 78% compared with 64% with operative repair. No difference was noted in recurrence rates, regardless of the timing of the leak onset (perioperative (p = 0.65) or acute (p = 0.44)). Of those patients who failed primary management, 4 (11%) eventually needed shunt placement for definitive management of their CSF leak. The overall mean difference in hospital related leak days was 3 days less in patients undergoing direct operative repair (p = 0.29). The overall mean difference in hospital related costs and actual charges were $7,856, and $1865.20 in favor of operative intervention as compared with lumbar drainage, though this was also not statistically significant.

Conclusion: Both lumbar drainage and operative repair are effective management strategies for treatment of CSF leaks after VS removal in the acute and perioperative periods. However, a subset of patients (11% in the present series) eventually needed shunt placement to treat their CSF leak due to pre-existing hydrocephalus. Further, an impressive trend toward lower hospital related costs and charges was found in patients undergoing operative management as primary treatment for acute CSF leaks.