Management of Intraoperative CSF Leaks during Endoscopic Endonasal Transsphenoidal Surgery: Methods and Efficacy

Background: Multiple methods have been described for managing intraoperative cerebrospinal fluid (CSF) leaks during endoscopic endonasal transsphenoidal surgery. We reviewed our experience with different methods of repair to ascertain risk factors for failed reconstruction.

Methods: We retrospectively reviewed all endoscopic endonasal procedures performed between 2011 and 2018 to identify those with intraoperative CSF leaks. Baseline demographics, reason for surgery, method of repair, lumbar drain usage, need for additional procedures, and complications were noted. Fisher’s exact test was used to assess for a relationship between failed repair and the following variables: use of a nasoseptal flap, inlay graft, onlay graft, fat graft, and lumbar drain with significance defined as p <  0.05.

Results: During the reviewed period, there were 27 procedures during which an intraoperative CSF leak was noted. The average follow-up was 18 ± 16 months. Thirteen of these procedures were in men, and fourteen in women with an average age of 58 years. Twenty-five were performed for pituitary adenomas, one was performed for a synovial sarcoma, and one was performed for a Rathke’s cleft cyst. A nasoseptal flap was utilized for 7 (26%) patients, an inlay graft was used in 26 (96%) of patients, an overlay graft was used for 23 (85%) of patients, a fat graft was used for 12 (44%), and a lumbar drain was used for 2 (7%) of patients. The most common inlays were dural substitute alone (n = 9, 33%), dural substitute and fat (n = 11, 41%), and fat alone (n = 5, 18%). The most common overlays were dural sealant alone (n = 9, 33%), acellular dermal matrix (n = 6, 22%), and dural substitute with dural sealant (n = 4, 11%). Persistent CSF leak was noted in 2 (7%) of cases. In one case, CSF diversion was attempted but failed and the patient ultimately required return to the OR for definitive repair with fat graft, acellular dermal matrix, and dural sealant. The other patient received a revision of his repair with acellular dermal matrix and fibrin glue as the initial and definitive treatment. One patient developed meningitis due to a persistent CSF leak. No additional complications were noted. None of the evaluated reconstruction variables were significantly associated with persistent CSF leak.

Conclusion: Repair of intraoperative CSF leaks during endoscopic endonasal surgery may be performed using multiple reconstructive methods including nasoseptal flaps, inlay, onlay, and fat grafts, as well as CSF diversion with a lumbar drain. These methods are efficacious in preventing persistent CSF leaks, though failure may require intraoperative revision of the reconstruction.