Significance of Empty Sella in the Management of Spontaneous CSF Leaks

February 5, 2009

Samar Fakhri, MD


In the traditional classification of cerebrospinal fluid (CSF) leaks, spontaneous (or idiopathic) rhinorrhea is included under the “normal-pressure” category. As such, the absence of a recognizable etiology is axiomatic to the diagnosis of a spontaneous CSF leak. Over the last decade, however, a number of published studies have described compelling demographic, clinical and radiographic characteristics common to most patients presenting with apparent spontaneous CSF leaks. The majority are obese, middle-aged women with an elevated Body Mass Index (BMI). The presence of an empty sella is a robust radiographic marker and is almost universally present in this group of patients, but not in patients with non-spontaneous leaks.1 These characteristics are largely similar to what is seen in Benign Intracranial Hypertension (BIH), raising the possibility that occult elevated CSF pressures may be an etiologic factor in the so-called spontaneous CSF leaks. In fact, mounting evidence now indicates that patients with apparent spontaneous CSF leaks represent a distinct clinical entity in which elevated ICP is a common clinical feature. In one study, Schlosser et al. reported that 8 of 11 patients (72%) with apparent idiopathic CSF rhinorrhea strictly fulfilled the modified Dandy criteria used by neuro-ophthalmologists to make the diagnosis of BIH, with the remaining 3 patients fulfilling most but not all criteria.2

From a pathophysiologic perspective, it has been postulated that chronically elevated CSF pressures enhance the pulsatile forces exerted on the skull base leading to attenuation and eventually erosion of inherently weak bony areas such as the lateral lamella of the cribriform plate or a laterally pneumatized recess of the sphenoid sinus. My experience corroborates these findings with the exception that none of my patients have had their defects located in the sphenoid sinus. The following case history summarizes the typical clinical presentation, salient features and management considerations of patients with apparent spontaneous CSF rhinorrhea.

A 44 year old woman was referred to our institution with a 1 year history of intermittent unilateral clear rhinorrhea. She reported severe headaches that improved when the rhinorrhea occurred. There was no history of accidental trauma or previous cranionasal surgery. She was obese and suffered from moderate hypertension. Office endoscopy demonstrated a small pulsatile mass just lateral to the anterior attachment of the middle turbinate.

The diagnosis of CSF fistula was confirmed with a positive beta-2 transferrin test. High resolution CT with a navigation protocol demonstrated a skull base defect in the supraorbital portion of the frontal recess/sinus with a sizable adjacent soft tissue density. MRI confirmed the presence of a frontal meningoencephalocele.

There was also a hyperintense T2 signal filling the sella, consistent with an empty sella.

At the time of surgical intervention, the opening CSF pressure was 16 cm H2O but increased to 32 cm H2O three days after successful endoscopic repair. The patient underwent endoscopic management with intraoperative navigation and intrathecal fluorescein injection. The encephalocele was fulgurated back to the intracranial cavity, the edges of the bony defect were identified and a layered repair with cartilage, bone and bovine acellular dermis was performed in an underlay fashion.

After defect localization, the management strategy should focus on intraoperative and postoperative considerations that maximize the outcome of surgical repair. The details of graft placement and selection have generated considerable controversy in the literature and are still largely dictated by the surgeon’s preference and comfort. We routinely employ a layered technique with bone and cartilage grafts as well as layers of bovine acellular dermis placed in the epidural space. This offers the advantage of reconstituting the layers of the skull base and thus may reduce the long-term failure rates.

Spontaneous CSF leaks have the highest failure and recurrence rates following surgical intervention. Recognition and long-term treatment of elevated ICP is therefore critical to the successful management of these patients. This may be achieved with diuretic therapy, such as acetazolamide, which is routinely employed in my patients. It should be mentioned that until today, there is no data to guide the timing, dosage and duration of such therapy, and long-term outcomes are still lacking. Less commonly, increased ICP may be lowered through a CSF diversion procedure such as ventriculoperitoneal shunting or lumbar peritoneal drain.

Endoscopic management of CSF leaks represents an early but elegant example of the evolution and effectiveness of transnasal endoscopic techniques in managing various sinonasal and skull base pathology. It also illustrates the multifaceted nature of our specialty where successful use of cutting edge technology often rests on understanding and manipulating basic physiologic principles.


  1. Schlosser RJ, Bolger WE. Significance of empty sella in cerebrospinal fluid leaks. Otolaryngol Head Neck Surg 2003;128:32-38.
  2. Zafereo ME, Fakhri S, Prayson R, et al. Esthesioneuroblastoma: 25-year experience at a single institution. Otolaryngol Head Neck Surg. 2008 Apr;138(4):452-8.