Persistent Cerebrospinal Fluid Leak after Endoscopic Endonasal Approach to the Posterior Cranial Fossa

 

Background Postoperative cerebrospinal fluid (CSF) leak is one of the main complications of the endoscopic endonasal approach (EEA) to the posterior fossa due to direct communication with arachnoid cisterns, higher CSF pressure, and more difficult defect reconstruction. A subset of patients require multiple repairs.

Objective Evaluate the characteristics of the population with persistent postoperative CSF leak after EEA to the posterior fossa.

Methods A retrospective chart review of patients with postoperative CSF leaks who underwent EEA to the posterior fossa between October 2010 and February 2017 was performed. Patients who required more than one consecutive CSF leak revision were included.

Results Nine patients matched our inclusion criteria. Mean follow-up interval was 17 months (4–45 months). There were eight males and one female in this group. The average patient age was 40.2 years (7–64 years) with one pediatric patient. Two patients were smokers, one had previous history of diabetes mellitus and one had chronic use of immunosuppressant drugs due to renal transplant. All patients were overweight or obese (5 obese) except the younger patient (mean BMI = 32.4, range: 15.5–54 kg/m²). Histopathology was chordoma in eight and epidermoid cyst in one; six had previous EEA for tumor resection, five performed in an outside hospital, and five had previous radiation. All patients had defects larger than 8 cm² (mean = 9.8 cm², range: 8.1–12.2 cm²). One patient did not have a vascularized intranasal flap option viable for reconstruction after tumor resection and was originally managed with fascia lata and fat graft, but ended up requiring a pericranial flap. Mean interval between first and second surgical CSF leak repair was 9 days (range: 4–15 days) and two developed a second episode of CSF leak even with lumbar drain (LD) or ventriculoperitoneal shunt (VPS). Median amount of consecutive episodes of CSF leak that required surgical repair was 3.4 (range: 2–7 consecutive surgical CSF leak repairs). All patients developed meningitis. Four patients developed flap necrosis, three identified in their first CSF leak repair and two in their second repair. One patient had necrosis of the nasoseptal flap (NSF) used in his original reconstruction, which was resected in his first revision and repaired with an inferior turbinate/lateral nasal wall flap (ITF), this also became necrotic and ended up requiring a pericranial flap (PCF).

All final reconstructions included at least one vascularized flap: three ITFs, three PCFs, two ITF combined with PCF, and one reused NSF. Four of the final reconstructions required endoscopic suturing of grafts/flaps to the basopharyngeal fascia and two patients had a ventricular shunt placed.

Conclusion Overweight or obese males being treated for recurrent chordoma are at increased risk for repetitive postoperative CSF leak. This is in part related to an apparent increased risk of flap necrosis. These large dural defects required a vascularized flap as part of their final reconstruction and almost half of them required endoscopic suturing. Further studies comparing a control group should be performed to confirm the influence of these findings in CSF leak recurrence.