Introduction

The term empty sella was first applied to an anatomic finding at autopsy by Busch in 1951. He found that in 40 patients with no known pituitary disease, the pituitary gland was severely flattened against the floor of the sella (5.6 percent of the series); the diaphragma sellae was restricted to a thin rim of tissue around a huge infundibular foramen. He also noted that an incomplete diaphragma might leave the pituitary gland intact but with its superior surface fully exposed and covered by arachnoid or might leave the pituitary indented and eccentrically placed within the sella. This latter condition resulted in a partially empty appearance. Busch pointed out that an empty sella was far more frequent in females than in males (34:6).

Little clinical attention was initially paid to Busch's interesting findings. Both radiologists and neurosurgeons usually regarded nontumoral sellar enlargement as reflecting pressure due to intrasellar arachnoidal cysts. However, Kaufman, in 1968, demonstrated that the "empty" sella was a manifestation of an enlarged intrasellar subarachnoid space. He noted that when the diaphragm a sellae was incomplete or incompetent, the subarachnoid space could expand, thus enlarging, deepening, and reshaping the sella.

The term empty sella syndrome appeared in two articles published in 1968. One of these papers applied the term to a patient with an enlarged sella that filled with air on pneumoencephalography; the patient had no known or demonstrable pituitary lesion. The other used the term for patients who had undergone prior surgery or radiation for proven pituitary tumors and who were later investigated for possible tumor recurrence when visual symptoms returned. An air study showed no tumor and an empty sella in each case.

Current usage is to call cases of sellar enlargement that are not related to a previously treated pituitary tumor primary empty sella syndrome, whereas those found following treatment of a pituitary tumor are denoted secondary empty sella syndrome. Abnormal dilation of the optic or infundibular recess as an isolated abnormality has also been reported. For our purposes, however, these will be considered as variants of the primary empty sella syndrome, discussed below. Secondary empty sella syndrome is discussed later.

Clinical Presentation

Although contemporary series suggest that the presence of an empty sella is fairly common (8 to 35 percent of the population), fewer than a third of individuals with the condition ever develop symptoms. Clinical studies of the symptomatic patients with primary empty sella syndrome have revealed a variety of associated problems, and it seems likely that several etiologic factors are involved in producing an empty sella. Despite this diversity, patients with primary empty sella syndrome are a surprisingly homogeneous group. Over 80 percent of these patients are women, more than 75 percent are obese, and the majority become symptomatic in the decade from age 40 to 49, with over 80 percent presenting between 30 and 59 years of age. An intrasellar problem is seldom suspected when the diagnostic studies are started.

Initial complaints fall into three categories: neurological, endocrinologic, and systemic. Headache is the most common presenting symptom, occurring in 50 to 80 percent of these patients. Other neurological complaints include memory loss, balance impairment, dizziness, seizures, and rhinorrhea. Papilledema, decreased visual acuity, and visual field loss are somewhat less common (10 percent). Endocrine complaints leading to investigation include amenorrhea, galactorrhea, loss of libido, and diabetes. Systemic complaints such as obesity or hypertension lead to radiologic studies, and sellar abnormalities are revealed.

By comparison, the rare presentation of the empty sella syndrome in children is usually marked by clinically significant endocrine abnormalities (45 to 75 percent of cases), craniofacial anomalies (30 percent), and visual complaints (as many as 35 percent). Taking note of the prevalence of endocrinologic findings and their frequent association with adverse perinatal events, Scire and colleagues showed that as many as 20 percent of children with precocious puberty have an empty sella. Smaller numbers of children with diabetes insipidus or with growth hormone deficiency will have an empty sella. Headache (13 percent) and nonspecific symptoms (4 percent) are much less common than in adults.

Radiologic Diagnosis

Plain skull film or tomographic measurements of the sella turcica should normally not exceed a length of 17 mm and a depth of 13 mm. In most patients with an empty sella, these measurements are greatly exceeded, and in a series of 20 patients the mean sellar volume was 2,77 ml. The primary empty sella syndrome may be responsible for up to 25 percent of cases of enlarged sella on plain films. However, volume measurements within normal limits do not rule out an empty sella, and a sella whose measured depth exceeds its anteroposterior length also calls for further investigation.

Plain roentgenographic changes typically include symmetrical ballooning of the sella, which still maintains a "closed" configuration. Although the floor may be considerably thinned, the clinoids are usually not attenuated. In the anteroposterior view, the sellar floor is symmetrical, in both position and bone thickness. Changes characteristic of pituitary tumor including posterior displacement and demineralization of the dorsum sellae and a double contour of the sellar floor; erosion of the entire sella can also be seen. These changes may be stable or progressive.

Computed tomography (CT) and magnetic resonance imaging (MRI) have replaced pneumoencephalography for the diagnosis of empty sella. On CT scans, an empty sella appears as a low-density abnormality in the pituitary fossa, often with erosion of the sellar floor. Following the intravenous infusion of a contrast medium. one can identify enhancement of the pituitary infundibulum and possibly the normal pituitary gland, which is usually displaced posteriorly and inferiorly. Coronal views are quite helpful. Before the advent of MRI, cases in which the diagnosis was in doubt were further subjected to CT metrizamide or air cisternography via a lumbar or cervical puncture. These techniques are now rarely necessary.

The key findings that distinguish an empty sella from a cystic mass on MRI are the isointensity of the fluid with cerebrospinal fluid (T1-weighted hypointensity and T2-weighted hyperintensity) and visualization of the infundibular stalk traversing the sellar cavity to the residual pituitary tissue. In cystic tumors or in cases of cystic degeneration of the pituitary, the infundibular stalk cannot be identified below the top of the tumor. This finding is also consistent with the diagnosis of arachnoid cyst.

Endocrine Studies

Clinical endocrine dysfunction is rare in adult patients with the empty sella syndrome; significant abnormalities are largely restricted to secondary amenorrhea, decreased libido, and loss of pituitary reserve. However, in some series the incidence of panhypopituitarism has been as high as 10 percent. Significantly altered pituitary stimulation tests are noted in 30 to 50 percent of these patients. The most commonly observed abnormality (30 percent of patients) has been deficient growth hormone (GH) secretion during stimulation tests. Abnormalities in the secretion of adrenocorticotropic hormone (ACTH) (11 percent of patients), luteinizing hormone (LH), and thyroid-stimulating hormone (TSH) (15 percent) have also been reported in association with an empty sella.

Hyperprolactinemia and intermittent increases in prolactin (PRL) levels have both been associated with the primary empty sella, and as many as 25 percent of women with an empty sella have elevated prolactin levels. The degree of hyperprolactinemia found in empty sella syndrome is moderate (usually less than 100 ng/ml). PRL values are more elevated with prolactinomas (nearly a 1:1 association with levels greater than 200 ng/m!). Patients with empty sellas have a normal PRL rise when stimulated with thyroid-releasing hormone (TRH). while patients with prolactinomas do not. Finally, the normal nocturnal peak PRL release is preserved with primary empty sella syndrome but is blunted with prolactinoma. These changes probably represent a "stalk effect" with interruption of the flow of dopamine (prolactin inhibitory factor).

Radiologic demonstration of an empty sella does not preclude the presence of a pituitary tumor. Functional pituitary tumors in "empty" sellas have been reported in acromegaly and Cushing's disease, and with prolactinoma. In fact, intrasellar arachnoid herniation appears to be more common with coexisting pituitary tumors than in the normal population. Intrasellar CSF extension has been observed in 14 percent of patients with acromegaly and in 17 percent of patients with prolactin-secreting tumors. This finding suggests that infarction of either the pituitary gland or of the tumor plays some role in the development of the empty sella in these patients. A similar mechanism has been well documented in patients with primary hypothyroidism who developed the empty sella syndrome following thyroxin-induced shrinkage of thyrotrope cell hyperplasia.

Normal pituitary function may persist despite extensive arachnoidal herniation with severe compression and flattening of the pituitary gland. However, cases of nontumoral hyperprolactinemia and of diabetes insipidus seen with primary empty sella syndrome suggest disruption of normal hypothalamic-pituitary pathways. It is possible that the arachnoidal herniation compresses the pituitary stalk at the rim of the diaphragma or upper portion of the bony sella. A similar mechanism has been proposed for certain visual problems associated with the secondary empty sella syndrome. In rare cases of empty sella syndrome following spontaneous infarction of a pituitary adenoma or in cases with Sheehan's syndrome, the precipitating process may itself be responsible for hypopituitarism.

Pathogenesis

The primary empty sella is evidently a multicausal syndrome, arising from several different pathogenetic processes. Factors to be considered in the formation of the primary empty sella are pituitary cysts, arachnoidal cysts, increased intracranial pressure, the prevalence of diaphragmatic incompetence, and ablative intrasellar events such as infarction or hemorrhage.

Developmental or degenerative incompetence of the diaphragma sellae is permissive but not always sufficient for the formation of the empty sella syndrome. Although the diaphragm a normally encircles the infundibulum snugly, Busch, Kaufman, and Bergland and colleagues have independently demonstrated diaphragma defects in 22 to 77 percent of patients. Most of these defects, however, do not result in herniation of the subarachnoid space, with only 5.1 to 6.7 percent of subjects in autopsy series showing empty sellas. Some other factor must be operative in these cases.

In his early description of the empty sella, Busch noted an unusually high incidence of cysts of the pars intermedia of the pituitary gland. Such cysts were usually microscopic, and were more common in female subjects. He postulated that these cysts might coalesce and rupture superiorly into the subarachnoid space. A deficient diaphragma sellae was needed to permit this rupture and thus create an "empty sella." This mechanism must be distinguished from the rare intrasellar epithelial cysts described by Fager and Carter. These epithelial cysts apparently arise from Rathke's pouch and are true mass lesions that may cause sellar destruction and optic nerve compression. They are not in free communication with the subarachnoid space.

Arachnoid cysts, either congenital or acquired, may extend from the suprasellar region downward into the sella if there is an incompetent diaphragm a sellae. Friedmann and Marguth described eight arachnoid cysts that had been operated on and suggested that with basal arachnoiditis a collateral pathway for the cerebrospinal fluid may result in sellar enlargement.

Although both pituitary and arachnoid cysts may occasionally provide an additional mechanism for the formation of the empty sella, conditions that cause increased intracranial pressure seem to be the culprit more often. Benign increased intracranial pressure (pseudotumor cerebri) also disproportionately affects obese females, and the similarity of this group of patients to those with empty sella probably reflects the partial causality of increased pressure. Weisburg and colleagues have demonstrated empty sellas in 10 percent of pseudotumor patients. Foley and Posner underscored this association, publishing 8 cases of empty sella associated with pseudotumor. Continuous intracranial pressure monitoring has demonstrated intermittent pressure increases in some patients with pseudotumor and empty sella, but others lack such rises. Severe systemic hypertension may play a role in such intermittent increases; it is present in as many as 31 percent of empty sella patients. Patients who lack such rises may have once had them, or may have an unrecognized predisposing condition such as treated primary hypothyroidism, unrecognized Sheehan's syndrome, or previous pituitary apoplexy. A careful history may disclose such a condition, which can provide the needed pressure or space for subarachnoid herniation in cases of diaphragmatic incompetence. Komatsu and colleagues and Mau and colleagues have independently demonstrated antipituitary auto antibodies in 33 to 75 percent of adults with the empty sella syndrome, suggesting another mechanism for pituitary destruction that may allow subarachnoid herniation.

Various other intracranial conditions that elevate cerebrospinal fluid pressure have been associated with the empty sella syndrome, including hydrocephalus, brain tumors, sarcoidosis, and Chiari malformations. Studies by Brismar and Bergstrand, using CT scanning, pneumoencephalography, and cisternography with isotope or metrizamide, have demonstrated impaired CSF circulation in most (>80 percent) patients with a primary empty sella. They pointed out the "obvious similarity" between the symptoms presented by these patients and those with normal-pressure hydrocephalus, including headache, moderate dementia, balance disturbances, and psychiatric symptoms. One of their patients benefited from a shunting procedure, and four others had temporary relief with lumbar punctures. More recently, Maira and colleagues have also demonstrated altered CSF dynamics in a large series of patients with the empty sella syndrome. These alterations correlate highly with prolactin abnormalities and symptomatology, both of which resolve with shunting.

Causes of empty sella syndrome:

In my opinion about the empty sella or partially empty sella syndrome:

1. The most common cause is the anatomical variation with wide aperture of the diaphragma sellae, through which the pituitary stalk is reaching the adenohypophysis. When it is wide, then the cardio-pulmonary pulsation with time make the sella wide and the pituitary gland becoming as the wall to the sella with the pituitary stalk hanging free. It is not a pathological finding and the person has no complain.
2. Arterial hypertension and increased intracranial pressure in the long-run can cause widening of the sellar cavity. Pseudotumor cerebri must be considered in the case of the presence of papilloedema.
3. Burned-out pituitary adenomas, after apoplexy and regression of the tumor from several causes, causing the already widened sella to become an empty sella.
4. Ruptured arachnoid cyst intrasellar localization.
5. Post radiation for several pathologic processes of intrasellar localization.
6. Post inflammatory processes, which caused in the past hypophysitis or other inflammatory process with subsequent scaring.
7. Miscellaneous

CSF Rhinorrhea

In 1968, Ommaya and colleagues reported two cases of CSF rhinorrhea associated with the empty sella syndrome, pointing out the relation between elevated CSF pressure and eventual rhinorrhea. They suggested that increased intracranial pressure might force open an anatomic defect not previously apparent, and thought that CSF leaks from the sella might result from persistence of the craniopharyngeal canal. Elevation of CSF pressure, combined with pulsations of the intrasellar arachnoid diverticulum, are likely causes of progressive sellar erosion and leakage.

Accurate figures on the incidence of CSF leakage in patients with primary empty sella syndrome are not readily available. Relatively few cases of trans-sellar leakage have been reported, in view of the high incidence of empty sellas in autopsy material. Weisberg and colleagues had two patients with rhinorrhea in their group of 25 empty sellas. Others suggest that CSF rhinorrhea may affect as many as 11.8 percent of those with an empty sella. In cases of nontraumatic CSF rhinorrhea, the sella is a far less common site of leakage than the cribriform plate.

Establishing the correct site of CSF rhinorrhea is of prime importance, and identification of an arachnoidal intrasellar herniation alone is not diagnostic; rather, it is necessary to demonstrate the actual site of leakage. Iohexol or metrizamide CT cisternography is currently thought to be more sensitive and specific than isotope scans.

Secondary Empty Sella Syndrome

A secondary empty sella results from herniation of the subarachnoid space into a vacant sella turcica following extirpation of pathologic sellar contents, whether by surgery or by radiation or by both. The usual lesion is an adenoma, and herniation may occur in the setting of either a congenitally deficient diaphragma or one that was weakened by tumor or therapeutic intervention. Secondary empty sella affects both sexes equally and usually presents with visual field deterioration several weeks to years after surgery or radiation. The pattern of visual loss is quite variable and nonspecific, and the loss may be aggravated by scarring of, and vascular injury to, the chiasm secondary to the original surgery and/or radiation. It is important to realize that the visual symptoms are worse and progress more rapidly than those seen with the primary empty sella syndrome. Radiographic findings are similar to those seen with the primary syndrome. Endocrine dysfunction is common but probably predates the intrasellar herniation. Rhinorrhea has been described, but it is probably largely the result of the original tumor and ensuing therapies.

Management: Medical and Surgical

Surgery is not indicated for the great majority of patients with an empty sella syndrome. The initial complaints of some of these patients are related to hypertension or obesity and require only medical treatment. The endocrinopathy associated with the primary empty sella syndrome is usually slight and rarely requires replacement therapy. Such a conservative approach is supported by Pompili and colleagues, who followed 12 women (mean age, 43 years) with a primary empty sella syndrome. In none of them did the disease progress, either clinically or radiographic ally . The one patient with a visual field deficit caused by the empty sella syndrome was ophthalmologically stable as well. However, overt diabetes mellitus developed in 15 percent of the cohort, and hypertension developed in twice that number. Together these diseases caused noncompressive, ischemic visual deficits in 2 of the 12 patients (15 percent). Furthermore, Beattie and colleagues have shown that as many as seven-eighths of empty sella syndrome patients suffer from coexisting glaucoma, which is readily treatable. Clearly, conservative treatment requires close medical and ophthalmologic follow-up.

Cases associated with benign intracranial hypertension (pseudotumor cerebri) are usually self-limited and best treated by non­surgical methods. An occasional patient may require lumboperitoneal CSF shunting to control papilledema or debilitating headache. Optic nerve fenestration has largely replaced subtemporal decompression for relief of optic nerve pressure. Patients with communi­cating hydrocephalus and elevated CSF pressure should undergo shunting without delay. When hydrocephalus is secondary to a tumor or Chiari malformation, the underlying pathologic process is usually treated first.

CSF rhinorrhea via an empty sella requires surgical intervention. Like other nontraumatic CSF leakage, this type of rhinorrhea seldom stops spontaneously. The first step in diagnosis is to determine whether a high-pressure or normal-pressure mechanism underlies the rhinorrhea. When the pressure is elevated by a tumor, diagnosis and surgical therapy is primarily directed at tumor diagnosis and removal. However, it is not enough to correct the cause; the leak must be repaired as well. When the leakage is due to hydrocephalus, shunting procedures may lead to the development of tension pneumocephalus. Transsphenoidal repair using fat and fascia, followed by a shunting procedure, may serve to avoid this problem. Occasionally, transcranial exploration is necessary if the site of leakage cannot be identified on preoperative studies.

When the intracranial pressure is normal, several treatment options are available. Continuous external lumbar subarachnoid drainage of CSF has been successfully employed to treat both postoperative and posttraumatic CSF fistulae. Continuous drainage is not without risk, and overdrainage may result in pneumocephalus. Several methods have been proposed to minimize this problem, including flow-regulated drainage.

Progressive ophthalmologic deterioration caused by chiasmal stretching or compression of associated vasculature may occasionally necessitate surgical intervention. This is especially true for the secondary empty sella syndrome. Chiasmapexy has been advocated by a few, but better results have been obtained using an extradural approach early after the onset of symptoms. By placing fat and fascia between the sellar floor and the sellar dura, one avoids injuring the intradural structures, which may include the downwardly displaced median eminence as well as the often enlarged intercavernous veins. Nagao and colleagues and Cybulski and colleagues have also performed chiasmapexy using transsphenoidal, intrasellar microballoon inflation with methyl methacrylate. Alternatively, Mortara and Norrell have described a lamina terminalisotomy which in theory vents the CSF pulsations and, in at least one case, resulted in successful treatment of a bitemporal hemianopsia. Lysis of adhesions and scar tissue surrounding chiasmal structures is to be avoided because little improvement can be expected and worsening has been reported.