Higher Recurrence Rate of Spontaneous Cerebrospinal Fluid Leak at Anterior Skull Base

Background Cerebrospinal fluid (CSF) leak of anterior skull base occurs in 2 to 3% of all cranial traumas and in 5 to 20% of all the previous skull base fractures. Their presentation usually occurs within 48 hours of the trauma and in most patients within a week. The longest interval described between the trauma and the presentation of CSF is up to ∼34 years. About 80% occur as a result of head trauma, 16% after surgery of paranasal sinuses, and 4% are spontaneous. The latter are rarely resolved with nonsurgical intervention. Bacterial meningitis is the most severe complication of CSF leaks in 2 to 50% of all cases which justifies the earlier diagnosis and treatment. The success rate is directly related to precise location of the site of CSF. In spontaneous cases, CSF leak usually occurs in the lateral wall of the sphenoidal sinus, especially when this is pneumatized toward the clinoid process below the maxillary nerve. The arachnoid granulation in this region usually has no associated venous drainage. When these vesicles filled with CSF pulse, this causes an erosion in bone wall and a herniation of the arachnoid into the sphenoidal sinus what precipitates fistula. Intracranial CSF leak approach could have a failure rate of even 40%. Endonasal endoscopic extracranial approach has a low rate of morbidity and a higher rate of success, of even 94% in the first proceeding and up to 100% with revision.

Methods From 2006 to 2011, clinical data of patients attended in the outpatient clinic of otorhinolaryngology in this period with CSF leak were collected prospectively. Information regarding the presence of prior or current meningitis episodes, etiology of the fistula (trauma, iatrogenic, and spontaneous), site (ethmoidal fovea, cribriform plate, and esfenoidal sinus), and surgery for closure of the defect and rate of success or recurrence were collected.

Results Of 65 patients, meningitis was present in 20% of the patients, being common more than an episode for patient, and one patient presented 12 episodes up to initiating his evaluation into our service. The length of follow-up was 5 to 10 years (average 7.8 years). The etiological distribution was trauma in 35%, spontaneous in 40%, and iatrogenic in 25%. In relation to site was ethmoidal in 32%, sphenoidal in 30%, cribriform plate in 33%, and frontal sinus in 8%. Those recurred cases were in sphenoid sinus. They presented recurrence in three cases (4.8%), all in sphenoid sinuses and only in spontaneous ones (p = 0.05). Rate of success was 95.4% in this cohort with endoscopic endonasal approach in first procedure with 100% with salvage surgery.

Conclusion Endoscopic endonasal approach had a high rate of success. Recurrence may have related to spontaneous cases and sphenoid sinus site.