Prevention and Treatment of CSF Leak in Endonasal Skull Base Surgery

Objective: Prevention and treatment of CSF-leak, water-tide closer of the bone defects after endonasal endoscopic approaches (EEA) to the skull base, especially in lesions with extrasellar extension is important stage of surgery.

Methods: A retrospective analyze of all transnasal procedures performed between 2007 and June 2016 was made. Patients with evidence of preoperative intra- and postoperative CSF-leak were included in this study. Several selected data concerning the type of pathologies, approaches, CSF-leak flow mode, method of reconstruction and associated complications were collected.

Results: A total 525 patients underwent surgery and 565 surgeries were performed. There were 433 (76.63%) patients with pituitary tumors, 21 (3.71%) with craniopharyngiomas, 34 (6.01%) with CSF-leak of different origin, 7 (1.23%) with meningiomas, 26(4.6%) with different skull base lesions, 4(0.7%) with RCC and 40 cases of repeat surgery. A 442 (84.19%) transsellar, 24 (4.57%) suprasellar, 32 (6.1%) transethmoidal, 6 (1.14%) transclival, 17 (3.23%) transsphenoidal and 5 (0.95%) combined approaches were performed. Patient’s age was between 16 and 77 years, with mean age of 49.01 years. The mean period of follow-up is 45.9 months.

A 152 patients with CSF-leak required a primary reconstruction. According to a CSF-leak mode 94 (61.84%) patients had low-flow, 51 (33.55%) moderate, and 7 (4.60%) high-flow CSF-leak. In 18 (11.84%) cases, reconstruction with spongostan, Vicryl patch and fibrin glue was made. Fat graft, a piece of fascia lata and fibrin glue was used in 23(15,13%) cases. Skull base reconstruction with vascular pedicle mucosal flap of the nasal septum (HB- flap) was used in 100 (65.78%) patients. In 2 (1.31%) patients a skull base ventral surface reconstruction with vascular pedicle pericranial flap was performed. And free mucosal flap of middle turbinate was applied in 9 (5.92%) patients. A postoperative cerebrospinal fluid leak occurred in 27 (17.76%) patients and was a reason of meningitis in 8 (1.9%) of them, all were cured. Two patients developed a spinal arachnoiditis due to lumbar drainage. Bothe was cured conservatively. In two patients lumbar drain was transferred to LP-shunt and one patient required a VP-shunting for hydrocephalus control. More often a CSF-leak occurs due to flap malposition. In one case CSF-leak occurred after radiotherapy and in two cases due to incorrect reconstruction with dead-space formation between in-lay and on-lay layers. In 2 (1.31%) patients a flap necrosis and flap perforation was a reason of recurrent CSF-leak respectively. Twenty patients (13.15%) with postoperative cerebrospinal fluid leak underwent a surgical reconstruction of CSF-fistula (five of them twice), in 7 (4.6%) cases CSF-leak resolved on lumbar drain. The mortality rate in series was 1.14%.

Conclusion: Usage of different methods for skull base reconstruction in endonasal surgery reduces risks of CSF-leak. Vascular pedicle flaps are more effective for water-tide closing in cases of extended approaches to the skull base. A learning curve impacts on the skull base surgery results and reconstruction’s effectiveness.