Cranial Base Reconstruction after Endoscopic Endonasal Surgery: Validation of an Algorithm in 832 Cases

Objective: Postoperative cerebrospinal fluid (CSF) leak remains a concern with endoscopic endonasal cranial base surgery. We present a reconstruction algorithm based on the type of intradural pathology and intraoperative CSF leak findings without the need for autologous fat or the use of temporary CSF diversion.

Methods: A common cranial base reconstruction algorithm for endoscopic endonasal surgery has been used at two institutions from 2005 to 2014. A retrospective review of a prospectively maintained database was performed. Data regarding case type, intraoperative CSF leak, repair method, and postoperative CSF leak were collected. All the cases were divided into the following three categories of intradural pathology: intrasellar, cranial base lesions requiring extended approaches, and spontaneous CSF leaks. Depending on the pathology and its anatomic location, as well as the intraoperative findings, grafts, biocompatible materials, and vascularized flaps were utilized. Pituitary lesions were further subcategorized into those without an intraoperative CSF leak, which were reconstructed with oxidized cellulose alone, and those with either an intraoperative CSF leak or patulous diaphragm following tumor resection, both of which were repaired with an inlay collagen allograft and polyethylene glycol sealant. Vascularized flaps were typically not utilized for these lesions. Cranial base lesions requiring extended endonasal approaches and dural transgression or resection were reconstructed with allograft or an inlay/onlay fascia lata graft for low-flow leaks and inlay/onlay fascia lata grafts reinforced with a vascularized nasoseptal flap for high-flow leaks. For spontaneous CSF leaks, allografts were used and reinforced with free mucosal or vascularized flaps depending on the defect size.

Results: The review included 832 consecutive patients with intradural pathology undergoing endoscopic endonasal surgery. With this cranial base reconstruction algorithm, the overall postoperative CSF leak was 2.3%. Of 550 patients with intradural sellar lesions, the majority of which were pituitary tumors, 10 (1.8%) patients had a postoperative CSF leak. In 253 patients with intradural cranial base lesions requiring an extended endonasal approach and having a high-flow intraoperative CSF leak, 7 (2.8%) demonstrated a postoperative leak. Overall, 29 spontaneous CSF fistulas were treated with one patient (3.4%) having a postoperative leak. The use of autologous fat and temporary CSF diversion was avoided in all the cases.

Conclusion: This algorithm presents a selective approach that can be safely used for reconstruction after endoscopic endonasal surgery, resulting in a postoperative CSF leak rate of 2.3%. Reconstruction with biocompatible materials and selected grafts without the need for autologous fat or temporary CSF diversion allows for effective cranial base repair with minimal nasal morbidity. In addition, minimal multilayer repairs and the avoidance of rigid allografts better allows for radiographic interpretation of follow-up imaging as well as reoperation, when necessary.