Spontaneous cerebrospinal fluid fistulae associated with empty sellae: surgical treatment and long-term results

To evaluate the accuracy of different diagnostic tools for cerebrospinal fluid fistulae (CFFs) associated with empty sellae and to assess the long-term results after surgical treatment.

Records were retrospectively reviewed for 38 of 44 patients who were treated for CFFs associated with empty sellae between 1972 and 1996. Glucose and radioisotope analyses of nasal fluids were used to establish the CFF diagnosis. Computed tomography, magnetic resonance imaging, and cisternography were used to localize the sites of the CFFs. Treatment was performed using three different approaches, i.e., subfrontal approach, transsphenoidal approach, and lumboperitoneal shunt placement.

Glucose and radioisotope analyses of nasal fluids confirmed CFFs in 97 and 100% of the cases, respectively. For localization of the sites, a leak in the sella was suggested in 32% of the cases using gammagraphy, in 50% using computed tomography or magnetic resonance imaging, and in 75% using computed tomographic contrast cisternography. After surgery, rhinorrhea ceased for 34 patients (89%), but 14 (41%) patients experienced recurrences of the rhinorrhea 6 months to 20 years after surgery (mean, 4.2 yr). Cure rates for the subfrontal approach, transsphenoidal approach, and shunt placement were 94, 85, and 86%, and the recurrence rates were 43, 33, and 50%, with mean times of 7.8, 4.2, and 0.8 years, respectively.

In this series, confirmation of CFFs was easy, although localization of the sites remained difficult. Recurrences were more numerous and occurred sooner after treatment with shunts; packing of the sella through a craniotomy achieved better and more stable results, although differences were not statistically significant.