2026 Proffered Presentations
S057: INCIDENCE OF CEREBROSPINAL FLUID LEAK FOLLOWING ENDOSCOPIC SURGERY FOR TUMORS OF THE VENTRAL SKULL BASE: SYSTEMATIC REVIEW AND META-ANALYSIS
Denise Sabac1; Michael Xie1; Michel Sourour2; Farshad Nassiri2; Eric Monteiro3; John de Almeida1; 1Department of Otolaryngology-Head and Neck Surgery, University Health Network, University of Toronto, Toronto, Ontario, Canada; 2Division of Neurosurgery, University Health Network, University of Toronto, Toronto, Ontario, Canada; 3Department of Otolaryngology-Head and Neck Surgery, Sinai Health Systems, University of Toronto, Toronto, Ontario, Canada
Introduction: Cerebrospinal fluid leaks (CSFL) represent a significant complication following skull base surgery. CSFL can occur anywhere along the skull base, even after adequate reconstruction. Unaddressed CSFL can predispose patients to life-threatening meningitis and pneumocephalus. Despite this risk, a comprehensive incidence measure has not been well established. While the incidence of postoperative CSFL following endoscopic skull base surgery has been reported, rates are inconsistent, ranging from 0% to over 30%. This discrepancy may be attributable to high heterogeneity across the literature for CSFL and one of the main factors is surgical subsite.
Objectives: This systematic review and meta-analysis aims to identify the incidence of CSFL following endoscopic surgery for ventral skull base tumors, stratified by subsite. Subsites were defined based on previously described sagittal and paramedian modules for expanded endonasal approaches.
Methods: This study adhered to PRISMA guidelines. Inclusion/exclusion: Papers reporting postoperative leak in adults (≥18 years old) undergoing endoscopic surgery for skull base neoplasms (benign or malignant) were included. Patients undergoing open or combined open/endoscopic surgery were excluded. Case series including fewer than six participants were excluded. Search strategy: A comprehensive search was performed in Embase, MEDLINE, CINAHL, and Web of Science databases following review by a librarian. The search was limited to studies published after 1990, following wider adoption of endoscopic endonasal surgery. Screening and data extraction: All papers were screened in duplicate, and data was extracted independently with a process to resolve differences. Demographic information and data on surgical approach, reconstruction method, pathology, subsite, incidence of leaks, flow rate, tumor volume/size, salvage setting, and use of lumbar drains were collected. Data synthesis and analysis: A meta-analysis of proportions using a random-effects model was implemented to calculate the pooled incidence of CSFL across studies and stratify by subsite. Common continuity correction was used to account for zero events. Heterogeneity was calculated using the I² statistic.
Results: The search retrieved 2,010 studies, and 168 met inclusion criteria. The pooled incidence was 5.97% (95% confidence interval [CI]: 5.08–6.85%, I² = 56.0%). By subsite, the pooled incidence was 5.48% (CI: 1.53–9.43%, I² = 34.4%) for sella, 4.70% (CI: 3.29–6.10%, I² = 63.2%) for pituitary tumors with extension beyond the sella, 6.77% (CI: 4.31–9.22%, I² = 56.4%) for suprasellar, 9.51% (CI: 3.38–15.7%, I² = 58.8%) for anterior cranial fossa, 8.90% (CI: 4.28–13.5%, I² = 0%) for clivus, 13.5% (CI: 6.36–20.7%, I² = 27.3%) for petroclival, 0.62% (CI: 0–5.42%, I² = 0%) for middle cranial fossa, 5.50% (CI: 0–13.8%, I² = 0%) for sphenoorbital, and 3.07% (CI: 0–14.1%, I² = 0%) for cavernous sinus. Other sites had fewer than three studies and were not included in the subsite meta-analysis. Seventy-two studies did not stratify CSFL by subsite and had a pooled leak rate of 6.30% (CI: 4.95–7.64%, I² = 41.3%).
Conclusion: Understanding the incidence of CSFL after endoscopic skull base surgery, stratified by subsite, establishes important prognostic data and provides a basis for quality indicators in future patient care and interventional research.