Establishing Hypercortisolism

Cushing’s syndrome, when suspected clinically, is confirmed by demonstrating hypercortisolism or characteristics of its effects on the normal functioning of the hypothalamic-pituitary-adrenal axis. Confirmation of excess cortisol production is made by one or more standard testing procedures. Most commonly today these tests include one or more of the following: serial 24-hour urine-free cortisol measurements, diurnal plasma cortisol levels, evening salivary cortisol levels to detect loss of diurnal rhythm of cortisol secretion, or the overnight dexamethasone suppression test.

Urine-free cortisol (UFC) measurements are assayed using a variety of techniques. The normal upper levels vary with the technique used and with the laboratory performing the assay (Fig. 20-2). Hypercortisolism is associated with loss of normal diurnal variation in cortisol secretion, which is demonstrated by obtaining morning (8 to 9 a.m.) and evening (11 to 12 p.m.) plasma cortisol levels (Fig. 20-3). Salivary cortisol levels are also reliably used for this and are well suited for outpatient screening of adult and pediatric patients for hypercortisolism.⁹ A study of more than 140 patients demonstrated a sensitivity of 93% and a specificity of 100% using this test to determine the presence of Cushing’s syndrome (Fig. 20-4).¹⁰ Because of its simplicity, the overnight low-dose (1 mg) dexamethasone suppression test is commonly used to screen patients for hypercortisolism (Fig. 20-5). In persons with a normal hypothalamic-pituitary-adrenal axis, AM cortisol levels are suppressed by the overnight low-dose dexamethasone suppression test (1.0 mg given the night before a morning [7 to 8 a.m.] cortisol measurement). A morning plasma cortisol level greater than 1.8 μg/dl after the bedtime (11 p.m.) administration of 1 mg of dexamethasone detects most patients with Cushing’s syndrome and justifies further diagnostic evaluation.¹¹

FIGURE 20-2 Twenty-four–hour excretion of urine-free cortisol to screen patients for Cushing’s syndrome. Comparison of 24-hour excretion of urine-free cortisol value in normal subjects and patients with confirmed Cushing’s syndrome.

(Modified from Loriaux DL, Cutler GB Jr. Diseases of the adrenal glands. In: Kohler PO, ed. Clinical Endocrinology. New York: Wiley; 1986:167-238, with permission.)⁶⁰

FIGURE 20-3 A. Diurnal rhythms of ACTH, β-endorphin, and cortisol in normal subjects. Plasma ACTH, β-endorphin, and cortisol concentrations in each of two (A and B) normal men who received blood sampling every 10 minutes for 24 hours. (Modified from Veldhuis JD, Iranmanesh A, Johnson ML, Lizarralde G. Amplitude, but not frequency, modulation of adrenocorticotropin secretory bursts gives rise to the nyctohemeral rhythm of the corticotropic axis in man. J Clin Endocrinol Metab. 1990;71:452-463.) B, Diurnal rhythms of plasma ACTH and cortisol levels in nine normal women (shaded area; ±95% confidence limits) and in five women with CD (mean ± standard deviation). L, lunch; D, dinner.

(Modified from Liu JH, Kazer RR, Rasmussen DD. Characterization of the twenty-four hour secretion patterns of adrenocorticotropin and cortisol in normal women and patients with Cushing’s disease. J Clin Endocrinol Metab. 1987;64:1027-1035.)

FIGURE 20-4 Salivary cortisols. Nighttime salivary cortisol values in healthy volunteers , patient controls , pseudo-Cushing patients , and patients with Cushing’s syndrome . Samples were collected at 12 midnight from inpatients and at bedtime (11 p.m. to 2 a.m.) from outpatients. To convert salivary cortisol nanograms per deciliter to nanomoles per liter, multiply by 0.0276.

(Modified from Papanicolaou DA, Mullen N, Kyrou I, et al. Nighttime salivary cortisol: a useful test for the diagnosis of Cushing’s syndrome. J Clin Endocrinol Metab. 2002;87:4515-4521.)

FIGURE 20-5 Detection of Cushing’s syndrome with the low-dose overnight dexamethasone suppression test. Plasma cortisol levels at 8 a.m. on successive days before and after 1 mg dexamethasone given orally at 11:00 p.m. in healthy subjects and patients with Cushing’s syndrome.

(Modified from Melby JC. Assessment of adrenocortical function. N Engl J Med. 1971;285:735-739.)

Differential Diagnosis of Hypercortisolism

Once excess cortisol production has been established, plasma ACTH levels are measured to distinguish between an ACTH-dependent and ACTH-independent etiology (Fig. 20-6; see also Fig. 20-1). With CD or ectopic ACTH secretion, the ACTH levels will be normal or elevated relative to the degree of glucocorticoid secretion. For this reason, these two entities are categorized as “ACTH-dependent” Cushing’s syndrome (see Table 20-1), in contrast to adrenal disease, in which plasma ACTH is low (<5 pg/ml) or undetectable (“ACTH-independent” Cushing’s syndrome, as the adrenal cortical cortisol secretion is autonomous).

FIGURE 20-6 Plasma ACTH levels distinguish patients with ACTH-independent (primary adrenal disease) and ACTH-dependent forms of Cushing’s syndrome. Plasma ACTH levels in 137 patients with Cushing’s syndrome.

(From Besser GM, Edwards CRW. Cushing’s syndrome. Clin Endocrinol Metab. 1972;1:451-490, with permission.)⁶⁴

At one time, the presence of Cushing’s syndrome and then differentiating CD from adrenal disease or ectopic ACTH secretion were examined with the 6-day dexamethasone suppression test, as described by Liddle.^12,13 The first 2 days of the test were used for measurement of basal cortisol secretion. In most patients with CD, cortisol secretion, an indirect measure of ACTH secretion by the pituitary gland or the tumor, is not suppressed during the 2 days of low-dose dexamethasone (0.5 mg every 6 hours for 48 hours), but 24-hour urinary cortisol secretion is suppressed to less than 10% of baseline values by 2 days of high-dose dexamethasone (2 mg every 6 hours for 48 hours). In contrast, high-dose dexamethasone fails to suppress cortisol secretion in most cases of ectopic ACTH secretion. Because of the difficulty in successfully completing this test today, it is now rarely used, but has been replaced with the high-dose overnight dexamethasone suppression test. For this test, 8 mg of dexamethasone is administered orally at 11 p.m. and a morning (7 to 8 a.m.) plasma cortisol measurement is obtained;¹⁴ for greatest diagnostic accuracy, suppression of morning serum cortisol of greater than 68% is required to assign a diagnosis of CD (Fig. 20-7).^14,15

FIGURE 20-7 High-dose (8 mg) overnight dexamethasone suppression test for the differential diagnosis of Cushing’s syndrome. Comparison of the high-dose overnight dexamethasone suppression test with the high-dose portion of the standard 6-day dexamethasone suppression test.

(Data from Tyrrell JB, Findling JW, Aron DC, Fitzgerald PA, Forsham PH. An overnight high-dose dexamethasone suppression test for rapid differential diagnosis of Cushing’s syndrome. Ann Intern Med. 1986;104:180-186.)

Another test available today to distinguish ectopic ACTH secretion from an ACTH-secreting pituitary tumor is the CRH stimulation test.^16–18 Since they are well-differentiated tumors derived from pituitary corticotrophs, most ACTH-secreting pituitary adenomas retain receptors for, and response to, CRH, whereas most ectopic tumors, tumors that are not derived from pituitary tissue, do not express receptors for CRH and do not respond to it (Fig. 20-8). The sensitivity and specificity of the test are optimal using ≥35% for the maximum ACTH response from the 15- or 30-minute samples to indicate CD (see Fig. 20-8B).¹⁸

FIGURE 20-8 A, CRH stimulation test for the differential diagnosis of Cushing’s syndrome. Plasma ACTH (top) and cortisol (bottom) responses to CRH in eight untreated patients with CD , six patients with Cushing’s syndrome due to ectopic ACTH secretion , and 10 human controls . (Modified from Chrousos GP, Schulte HM, Oldfield EH, Gold PW, Cutler GB Jr, Loriaux DL. The corticotropin-releasing factor stimulation test. An aid in the evaluation of patients with Cushing’s syndrome. N Engl J Med. 1984;310:622-626.)B, Responses of plasma ACTH and cortisol to intravenous administration of ovine CRH in patients with CD and ectopic ACTH secretion. ACTH responses are expressed as the percent change in mean ACTH concentration 15 and 30 min after CRH from the basal value 1 and 5 min before the injection. Dashed line indicates a response of 35%. Cortisol responses are expressed as the percent change in mean cortisol concentration 30 and 45 min after CRH from the basal value 1 and 5 min before the injection. Dashed line indicates a response of 20%.

(Modified from Nieman LK, Oldfield EH, Wesley R, et al. A simplified morning ovine corticotropin-releasing hormone stimulation test for the differential diagnosis of adrenocorticotropin-dependent Cushing’s syndrome. J Clin Endocrinol Metab. 1993;77:1308-1312.)

If ACTH-dependent hypercortisolism is established, a sella MRI with and without contrast is obtained (see below). In many patients with CD, a high-resolution sella MRI will demonstrate the presence and location of a pituitary tumor. However, a negative sella MRI does not rule out an adenoma, as the false-negative rate using the standard T1-weighted spin echo after contrast enhancement in CD is as high as 50% at some centers.

If the results of the high-dose dexamethasone suppression test and the CRH stimulation test are consistent with CD and the pituitary MRI reveals a definite adenoma, no further diagnostic testing is necessary. However, if either of these provocative endocrine tests is inconsistent with CD, inferior petrosal sinus sampling is performed.

The test with the greatest diagnostic accuracy for the differential diagnosis of CD versus ectopic ACTH syndrome is bilateral simultaneous inferior petrosal sinus sampling performed with and without intravenous CRH administration.¹⁹ The test is performed by placement of catheters with their tips in the inferior petrosal sinuses and in a peripheral vein (Fig. 20-9A), and then obtaining serial, simultaneous samples for central and peripheral plasma ACTH concentrations at 2 and 0 minutes before and at 3, 5, and 10 minutes after intravenous CRH administration (1 μg/kg body weight). IPSS is only used in patients with confirmed hypercortisolism, as the test cannot discriminate between normal subjects and patients with CD. Further, since this is an invasive procedure with rare but serious associated risks,²⁰ it is generally used in patients in whom the results of provocative endocrine testing to distinguish ectopic ACTH secretion from CD are conflicting or equivocal and in instances in which the sella MRI is negative. In this test the levels of ACTH in the primary venous drainage of the pituitary, the inferior petrosal sinuses, are compared to simultaneous ACTH measurements in the peripheral blood. A peak ratio of 2:1 petrosal:peripheral during baseline (before CRH) or 3:1 before or after CRH indicates a pituitary source of the excess ACTH, that is, CD.¹⁹ The sensitivity of the test is increased by sampling after administration of CRH, which, because it stimulates a pituitary adenoma to secrete ACTH, enhances the ACTH concentration differential between the central (inferior petrosal sinus) and the peripheral blood (Fig. 20-9B).¹⁹ It originally seemed that the procedure had a diagnostic accuracy 100%. However, reports of diagnostic errors, almost all false negative, have appeared, although in our experience the diagnostic accuracy approaches 100%. The accuracy of the test also relies upon successful placement of the catheter tips in the petrosal sinus bilaterally, since a false negative result may arise if the blood from both inferior petrosal sinuses cannot be catheterized successfully. Thus, venography is performed to ensure correct catheter placement and to evaluate the venous anatomy. A hypoplastic or anomalous inferior petrosal sinus in 0.8% of 501 patients was associated with false-negative results in patients with proven CD.²¹ It was briefly thought that comparison of petrosal sinus ACTH from the right to left sides would accurately indicate the side of the pituitary in which a small pituitary tumor was located, permitting a more focused search for it at surgery, or permitting removal of the half of the pituitary containing an adenoma that was too small to identify despite a thorough search of the gland during surgery.²² However, more experience with the technique for lateralization indicated that the lateralization accuracy is only about 70%, and even less so in pediatric patients,²³ compromising its usefulness as a localizing measure during surgery.¹⁹

FIGURE 20-9 A, Anatomy and catheter placement in bilateral simultaneous blood sampling of the inferior petrosal sinuses. Confluent pituitary veins empty laterally into the cavernous sinuses, which drain into the inferior petrosal sinuses. B, Venous sampling during inferior petrosal venous sampling with ACTH levels from a patient with CD (left) and a patient with a bronchial carcinoid and ectopic ACTH secretion (right). C, Bilateral inferior petrosal vein sampling in the differential diagnosis of Cushing’s syndrome. Maximum ratio of ACTH concentration from one of the inferior petrosal sinuses to the simultaneous peripheral venous ACTH concentration in patients with Cushing’s syndrome in basal samples (left), and in basal and CRH-stimulated samples (right). During basal sampling, the maximum IPS:P ACTH ratio was ≥2.0 in 205 of 215 patients with confirmed CD, but was <2.0 in all patients with ectopic ACTH syndrome or primary adrenal disease. All patients with CD who received CRH had maximum IPS:P ACTH ratios of ≥3.0, whereas all patients with ectopic ACTH syndrome had IPS:P ratios of <3.0. The asterisks represent five patients with primary adrenal disease in whom ACTH was undetectable in the peripheral blood before and after CRH administration.

(A and C from Oldfield EH, Doppman JL, Nieman LK, et al. Bilateral inferior petrosal sinus sampling with and without corticotropin-releasing hormone for the differential diagnosis of Cushing’s syndrome. N Engl J Med. 1991;325:897-905.)

Imaging

Sella magnetic resonance imaging (MRI) is the imaging procedure of choice for detecting and localizing the pituitary adenoma in patients with CD. MRI should be performed with and without contrast, as the adenomas typically have decreased enhancement compared to the normal gland (Fig. 20-10). The resolution of a 1.5-T magnet may reveal tumors as small as 3 mm in diameter. MRI provides other important anatomical information for the surgeon: aeration of the sphenoid, parasellar anatomy, location of the carotid arteries or coexisting aneurysms, extent of supra- or para-sellar extension of an adenoma, and ectopic parasellar tumors. Recently the spoiled gradient recalled acquisition (SPGR) technique, used with 1-mm nonoverlapping slices was shown to be more sensitive than the conventional spin echo approach (sensitivity 80% vs. 49%), a finding also true in pediatric CD.^24,25 However, the incidence of false positives was also higher (4% vs. 2%). Since the sensitivity of the conventional MRI techniques (spin echo) for CD is only 50% to 75%,^26–29 many patients have a negative MRI. Furthermore, MRI is not always available or possible, as in patients with an MRI incompatible cardiac pacemaker or a morbidly obese patient who cannot be accommodated by the scanner. In these patients, a sella computerized tomography image (CT) with and without contrast may demonstrate a tumor, but CT is less sensitive than MRI. Furthermore, since the specificity of MRI is not 100% (MRI abnormalities consistent with adenomas occur in the pituitary gland in 10% of normal volunteers³⁰ and incidental adenomas are found in the pituitary gland in about 5% to 20% of subjects in unselected autopsy studies^31,32), the results of MRI scanning must be confirmed by endocrinologic testing before surgery. In general, CD endocrine testing provides the diagnosis and CT localizes the adenoma within the pituitary and defines the anatomy in its vicinity.

FIGURE 20-10 Magnetic resonance imaging of the pituitary in a patient with a 4-mm adenoma in the right half of the anterior lobe. This image was acquired before (left) and immediately after (right) infusion of gadolinium-DTPA (T1-weighted MRI, TR 500, TE 20).

(From Doppman JL, Frank JA, Dwyer AJ, et al. Gadolinium DTPA enhanced imaging of ACTH-secreting microadenomas of the pituitary gland. Correlation of MR appearance with surgical findings. J Comput Assist Tomogr. 1988;12:728-735.)

Surgery

Trans-sphenoidal microsurgery is the treatment of choice for CD. Identification of the adenoma and selective adenomectomy provides remission of hypercortisolism in most patients with an adenoma that is contained within the anterior lobe and that is large enough to be detected by MRI. Surgery for these tumors is similar to the procedures that have been described for pituitary tumors in general.

However, many tumors that cause CD are small, frequently so small that they are not identified on the preoperative MRI. In these patients the first task for the surgeon is to locate the adenoma. A wide exposure of the pituitary gland permits visualization of the anterior lobe from one edge to the other. This requires removal of the bone of the anterior face of the sella until the medial portion of the cavernous sinus is visible bilaterally. The pseudocapsule of a microadenoma is almost always a grey-white or gray-yellow color. Because coagulation of the dura will discolor a circumscribed area of the pituitary surface, producing a local white region that can mimic an adenoma just beneath the site of dural coagulation, bipolar cautery of the dura is avoided. A wide cruciate dural opening extending to the lateral corners of the exposure, to the junction of the circular sinus with the cavernous sinus superiorly and to the medial wall of the cavernous sinus inferiorly, while avoiding entry into the capsule of the pituitary gland, allows visualization of most of the anterior surface of the anterior lobe. In many cases, an adenoma will be seen during careful inspection of the gland’s surface while using the operating microscope to look for a focal mass or discoloration on the surface of the gland. Localization of adenomas with intraoperative ultrasound, using a specially designed probe, has also proven useful in patients with negative MRI scans.³³ When it is used, intraoperative ultrasound is obtained after the bone removal is complete, but before the dura is opened, and with a bloodless field. However, in our experience the most reliable clue to the identification of a microadenoma is visualization of its pseudocapsule. This is especially important for the small tumors that are buried in the gland.

When the location of an adenoma is identified by inspection of the surface of the gland, the pituitary capsule is sharply incised just beyond the visible margin of the tumor pseudocapsule, which is a margin of compressed normal gland surrounding the adenoma (Fig. 20-11). If the adenoma is not identified by examining the pituitary capsule anteriorly and inspecting the lateral surfaces of the anterior lobe, a series of incisions is made vertically in the gland at intervals of 1.5 to 2 mm, each carried deeper in increments until either the pseudocapsule of the tumor is identified or the juncture of the anterior and posterior lobes is reached (Fig. 20-12). By identification of the interface between the pseudocapsule of the adenoma and the surrounding normal gland, the adenoma is discretely dissected from the normal pituitary, ideally without entering the soft body of the adenoma. In this way, total resection is commonly performed and lasting remission is achieved.³⁴ If the tumor breaches the capsule of the pituitary gland, careful inspection of the contiguous dura is performed to ensure that dural invasion by the tumor is not overlooked. This is particularly important laterally along the medial wall of the cavernous sinus, where even small tumors may spread into the adjacent dura and cavernous sinus.

FIGURE 20-11 A, Schematic drawings showing exposure of the pituitary. The initial opening of the anterior face of the sella turcica is performed using a drill with a 3- to 4-mm burr. A 2-mm Kerrison ronguer with a thin distal lip (0.75 mm) is used to remove the remaining thin layer of bone of the anterior sella face in small increments. These small pieces of bone can be flushed away with a brief pulse of saline irrigation, which is performed by the assistant, and suctioned away, permitting removal of the pieces without removing instruments from the working region. The bone is removed laterally until at least 2 mm of the most medial aspect of each cavernous sinus is visible. Superiorly bone removal extends to the tuberculum sellae. Inferiorly, a disc dissector is used to separate gently the dura from the bone of the sella floor after the anterior face of the sella has been removed. The sella floor is then removed with a small pituitary rongeur. Oozing bone margins are covered with a thin layer of bone wax and compressed into the bone interstices by using manipulation of a cotton pledget with the forceps tips. Because of the very low pressure in the cavernous sinus and the smaller dural veins draining into it, the same technique, or placement of small pieces of Gelfoam soaked in thrombin, can be used successfully for any site that slowly oozes blood from the layer between the dura and bone at the margins of the bone removal. At this stage the operative field should be bloodless and should expose the entire anterior sella dura and most of the inferior dura covering the pituitary. Much of the success of the surgery depends on achieving a wide and completely bloodless exposure. Inspection of the dura reveals any region of invasion of the anterior or inferior dura and, in cases of small tumors that reach the anterior surface of the pituitary gland, careful examination of the dural surface may provide clues to the site of the tumor, such as a slight local protuberance or a focal region of dark or light color. Because coagulation of the dura tends to glue it to the underlying pituitary capsule (impeding the capacity to open the dura sharply while leaving the pituitary capsule intact) and because it produces a region of white discoloration of the pituitary capsule and the contiguous region of the pituitary gland immediately beneath it (discoloration that may be misleading when searching for a small microadenoma), coagulation of the anterior or inferior dura is avoided during the exposure. All suction from this point is through a 0.5-inch square cotton pledget onto which the suction tip is placed close to the margin. The dura is opened using a No. 15 scalpel with care to take the incision completely through the dura, but not to enter the capsule of the pituitary gland. To achieve the widest available exposure, the superior margins of the incisions are extended laterally to reach the interface of the circular sinus and the medial margin of the cavernous sinus (upper inset). The cavernous sinus is often intentionally entered along the superior and lateral margin of the exposure, where the lateral aspect of the circular sinus meets the medial aspect of the superior portion of the cavernous sinus. Low-pressure venous bleeding from this opening is controlled by simply plugging the opening with a small piece of Gelfoam moistened in thrombin. To achieve a wide exposure laterally and inferiorly, the medial inferior region of the cavernous sinus is frequently opened during the lateral most extent of the incision and bleeding is controlled in a similar fashion. The dura is gently and cleanly separated from the pituitary capsule by using a disc dissector in the tissue plane between these layers. This exposes the entire anterior surface of the pituitary gland with an intact pituitary capsule in a bloodless field. Similar to the pia mater, the capsule of the pituitary gland (lower inset) is a strong layer despite its translucent nature and, even though it is visible histologically, like the pia mater, the capsule is invisible to the surgeon. It must always be cut sharply; blunt dissection will not open it and will only transfer pressure to the interior of the gland and to the tumor, potentially spilling the partially liquefied center of the tumor prematurely. B and C, The use of the surgical capsule for selective excision of small adenomas that are visible on MR imaging. When dealing with tumors within the gland that are visible on MR images, a curvilinear incision is made through the pituitary capsule just beyond the point at which the most superficial dome of the tumor reaches, or comes closest to reaching, the surface of the gland (B, left). Thus, this initial incision is not made directly into the tumor, as is common practice. This permits a thin layer of normal gland to be passed through before reaching the surgical capsule of the adenoma (left inset), allowing easy identification of the surgical capsule and the creation of a surgical plane of dissection at the margin of the tumor, at the interface of the normal gland and the surgical capsule of the adenoma. This interface between the adenoma and gland is further defined using the tips of the bipolar forceps in a series of movements parallel to the surface of the adenoma and in the crevice between the gland and adenoma (B, right and center and right insets). If the correct tissue plane is used, the dissection is unimpeded and the gray–white surface of the adenoma is spherical and smooth. After this interface has been clearly defined, gentle dissection of the interface between the adenoma and gland is continued following the curvilinear margin of the adenoma, and with further incisions in the capsule of the pituitary gland just beyond the tumor margin to release tension on the tumor as the dissection proceeds. A small Hardy sucker (2-mm tip) on the margin of a cotton pad is used to provide separation of the interface between the gland and the adenoma for dissection and for sponging the small amount of bleeding out of the field of view. Small pieces of Gelfoam soaked in thrombin help preserve the surgical space along the dissected margins while dissection takes place at other sites around the tumor margin (not shown here). After the most superficial portion of the tumor has been defined circumferentially, the deeper adenoma margins are defined and dissected in a similar fashion (C); the posterior margin of the adenoma often requires dissection using a disc dissector and a small and/or medium ring curette (C, left and center inset). This is performed gently; very little pressure is required if the correct tissue plane is being used, and most of the limited pressure that is applied is directed more toward the gland than the adenoma. After the margins of the tumor have been completely dissected, to prevent rupture of the tumor the last remaining connection between the pseudocapsule of the specimen and the pituitary capsule is grasped with a small cup forceps and the tumor is removed (C, right). In cases of tumors 8 to 10 mm or less in diameter, the entire tumor can usually be shelled out of its bed in the anterior lobe as an intact specimen. Successful and complete removal leaves a smoothly lined hemispherical tissue void in the anterior lobe.

(From Oldfield EH, Vortmeyer AO. Development of a histological pseudocapsule and its use as a surgical capsule in the excision of pituitary tumors. J Neurosurg. 2006;104:7-19.)

FIGURE 20-12 Exploration of the pituitary gland in patients with an endocrine-active adenoma whose imaging studies appear normal. The goal of the exploration is to find and identify the distinct encapsulated margin of the adenoma. Success depends on beginning with a widely exposed, bloodless surgical field (A). Coagulation of the anterior and inferior sella dura is avoided during the exposure and dural opening because it would produce a white area on the surface of the underlying gland that may falsely suggest the site of the adenoma. After widely opening the dura, which provides exposure of an intact pituitary capsule covering the anterior surface of the gland, and after exposure of the extreme lateral margins of the anterior surface, the surfaces are carefully inspected for regions of focal discoloration. The adenoma usually appears to be gray-blue or yellow-white, and can be identified against the background of the anterior lobe surface whose color is orange-pink. B, If no tumor is seen, the lateral surfaces of the anterior lobe are then inspected. For this the lateral dural wall of the sella (the medial wall of the cavernous sinus) is separated from the pituitary capsule by gently passing a disc dissector between these two layers from top to bottom. The space produced provides room for dissection of the interface between the lateral pituitary capsule and the dural wall with the closed tips of a fine-tipped bipolar forceps (B). Dissection is initially superficial and then progresses in stages to deeper levels until the posterior sella has been reached. After these two tissues have been separated, small pieces of Gelfoam are packed into the intervening space to rotate the lateral surface anteriorly and to gently displace it medially into the surgeon’s direct view. After both lateral surfaces are exposed and examined in this fashion, the inferior surface of the gland is separated from the dura. (This is the site at which these two layers—the surface of the pituitary and the dura mater—are most tenaciously attached to each other.) If a tumor can be identified from inspection of the surface, it is removed as described in the text and the legend to Fig. 20-11. If no tumor is identified on inspection of the superficial gland, a series of vertical incisions is then made (B), each of which begins 1 to 2 mm below the superior edge of the pituitary exposure and is directed downward, initially to a depth of only approximately 1 mm, and then in stages deeper through the anterior lobe until either the intermediate lobe or the glistening white anterior surface of the posterior lobe is reached. Because the pituitary blood supply and the delivery of hypothalamic trophic factors to the pituitary is oriented vertically, vertical incisions should be less likely to cause an infarction in a portion of the pituitary or to isolate the gland from its hypothalamic regulation. A distinct tissue capsule is the primary object of the search, with the intent to identify the margin of the tumor before entering it and spilling its contents, by using the surgical capsule of the adenoma (B and C). When the surgical capsule of the adenoma is identified, the tumor is removed using dissection along the interface of the adenoma and the normal gland, as described in Fig. 20-11 (A and B). In cases of tumors lying in the posterior portion of the anterior lobe, a 2-mm-wide slice of the anterior lobe may be removed to provide space for dissection and removal of the deep microadenoma (not shown).

(From Oldfield EH, Vortmeyer AO. Development of a histological pseudocapsule and its use as a surgical capsule in the excision of pituitary tumors. J Neurosurg. 2006;104:7-19.)

ACTH-producing tumors may also be found in the posterior lobe,³⁵ so careful inspection of the gland back to the neurohypophysis should be preformed when no tumor is encountered anteriorly.

In the rare instance when the ACTH-secreting adenoma is located outside of the adenohypophysis, such as in a suprasellar (Fig. 20-13) or parasellar location,³⁶ special techniques are required. For tumors of the pituitary stalk in the suprasellar cistern, an extended trans-sphenoidal approach may be used.³⁷ With this technique, during the standard trans-septal approach, bone removal is extended rostrally over the planum sphenoidale. The dura anterior to the gland is opened in the midline to expose the suprasellar cistern (Fig. 20-14). Bleeding from the opened circular sinus is stopped by packing it with small pieces of Gelfoam or with bipolar coagulation. Direct visual access of the pituitary stalk is thus provided.

FIGURE 20-13 Adenoma in pituitary stalk. Magnetic resonance T1-weighted images revealing a 7-mm adenoma arising high in the pituitary stalk, just beneath the optic chiasm (left, sagittal view; right, coronal view) after contrast enhancement with gadolinium DTPA. Note enhancement of the anterior lobe and stalk, but not the tumor.

(From Mason RB, Nieman LK, Doppman JL, et al. Selective excision of adenomas originating in or extending into the pituitary stalk with preservation of pituitary function. J Neurosurg. 1997;87:343-351.)

FIGURE 20-14 Extended trans-sphenoidal technique for removing suprasellar tumor of the pituitary stalk. Drawings depict the routine wide exposure of the anterior surface of the sella and bone removal to expose the medial portion of the cavernous sinus bilaterally. The posterior portion of the planum sphenoidale (the posterior 4–6 mm) is removed by first drilling with a rough diamond burr until the plate of bone was paper thin, and then using a 2-mm-thin–footplate cervical Kerrison rongeur. Removal of the planum sphenoidale aids access to the suprasellar cistern, the pituitary stalk, and the superior surface of the gland. The dura mater covering the anterior pituitary surface is opened widely. Bilateral parasagittal incisions are made 8 to 10 mm apart in the dura overlying the planum sphenoidale. A transverse incision just above the anterior portion of the circular sinus is then used to connect the parasagittal incisions. The resulting dural flap is opened posteriorly while, initially, preserving the intact arachnoid. After entering the suprasellar cistern, the superior hypophyseal artery is identified and care is taken not to injure it. The exposed diaphragma sella is incised in the midline, in an anteroposterior direction, to reach the stalk and the supradiaphragmatic tumor. A small piece of Gelfoam or cottonoid is placed superiorly in the subarachnoid space to prevent passage of blood into the cerebrospinal fluid. Characteristic vertical striations produced by the vertical course of the surface blood vessels permit identification of the stalk. A sharp incision is made in the pia with a No. 15 scalpel at the junction of the capsule of the tumor with the pituitary stalk and, when appropriate, in the superior surface of the anterior lobe at the margin of the adenoma. The adenoma is then resected using standard microsurgical technique.

(From Mason RB, Nieman LK, Doppman JL, et al. Selective excision of adenomas originating in or extending into the pituitary stalk with preservation of pituitary function. J Neurosurg. 1997;87:343-351.)

Endonasal or endoscopic approaches to the sella are being examined, although their utility for resection of endocrine-active adenomas, such as CD, remains to be established.

Postoperative Assessment

The chronic hypercortisolism associated with CD suppresses the hypothalamic CRH-producing cells and the secretion of ACTH by the normal corticotrophs, but does not fully suppress ACTH secretion by the tumor. In most patients who are cured by surgery it takes several months for the hypothalamic-pituitary-adrenal axis to recover and the patient suffers from hypocortisolism and requires glucocorticoid replacement therapy for the 6 to 24 months required for recovery of the hypothalamic neuron, the normal pituitary corticotroph, and the adrenal cortex and return of normal endogenous circadian cortisol secretion (Fig. 20-15). Therefore, complete removal of the abnormal source of ACTH should produce hypocortisolism in the immediate postoperative interval and for several months.³⁸ Thus, in almost all instances of successful surgery for CD, the patient will be hypocortisolemic (a morning cortisol of less than 5 μg/dl and 24-hour urine-free cortisol secretion of less than 20 μg). Because of this, postoperative cortisol replacement must be provided. There are several protocols for postoperative cortisol replacement therapy. One approach used in our practices is to administer 0.5 mg of dexamethasone every 6 hours for 36 hours. The dexamethasone is then discontinued and daily AM cortisol and 24-hour urine cortisol are obtained for 3 consecutive days. This permits immediate determination if surgery has been successful in eliminating hypercortisolism and provides the potential for immediate early reoperation, if indicated, which is successful in many patients.³⁹ Others use postoperative low-dose dexamethasone suppression testing as follows: on the first postoperative day, the patient receives 50 mg of hydrocortisone intravenously in the morning and 25 mg in the evening. On the second day, they are given 50 mg of hydrocortisone in the morning and 1 mg of dexamethasone in the evening at 10 p.m. Blood is withdrawn from the patient at 8 a.m. on the third postoperative morning and a fasting serum cortisol level is measured. With this approach patients with a morning plasma cortisol greater than 2 μg/dl will have a high likelihood of having recurrent disease: 100% of patients in the report by Arnott et al.,⁴⁰ and in the larger experience of Chen et al.⁴¹ On the other hand, 93% of patients with an AM cortisol value of 2 μg/dl or less still had sustained remission of CD at 5 years (Fig. 20-16).⁴¹ Many authors prefer either to provide cortisol replacement therapy and to test the patient’s cortisol secretion several weeks after hospital discharge, or to provide no cortisol replacement and sample serum cortisol levels every 6 hours after surgery.^42–44 Most cured patients have serum cortisol drop below 2 μg/dl within the first 48 hours after surgery.

FIGURE 20-15 Recovery of the hypothalamic-pituitary-adrenal axis from chronic hypercortisolism. Nine patients with Cushing’s syndrome (six patients with CD, two with adrenal adenomas, and one with ectopic ACTH syndrome) were tested longitudinally for 12 months after surgical correction of Cushing’s syndrome with the 1-hour ACTH test (lower panel) and the oCRH test (upper and middle panels). The times at the top of the figure indicate the interval between the testing and surgery and the number of patients studied at each time point is shown at the bottom. The shaded areas represent the mean (± standard deviation) responses in normal subjects.

(From Avgerinos PC, Chrousos GP, Nieman LK, Oldfield EH, Loriaux DL, Cutler GB Jr. The corticotropin-releasing hormone test in the postoperative evaluation of patients with Cushing’s syndrome. J Clin Endocrinol Metab. 1987;65:906-913.)

FIGURE 20-16 Postoperative remission of CD predicted by the overnight low dose dexamethasone suppression test. Kaplan-Meier graph compares recurrence-free survival for patients with AM plasma cortisol levels lower than 3 μg/dl (solid line) or 3 to 8 μg/dl (broken line) on the third postoperative day after receiving 1 mg of dexamethasone at 10 to 11 p.m. the previous evening.

(Modified from Chen JCT, Amar AP, Choi SH, et al. Trans-sphenoidal microsurgical treatment of Cushing’s disease: postoperative assessment of surgical efficacy using overnight low-dose dexamethasone suppression test. J Neurosurg. 2003; 88:34-37.)

Without glucocorticoid replacement, the patient may be symptomatic from an Addisonian state and will typically complain of symptoms of lethargy, anorexia, and abdominal discomfort or nausea. As the patient’s blood pressure will be normal or even low postoperatively, preoperative medications used to treat hypertension are withheld until it is determined that they are required after surgery. Similarly, since glycemic control may be restored after surgery, preoperative diabetic medications are withheld, and blood glucose is carefully monitored in the immediate postoperative interval. After discharge twice daily physiologic cortisol replacement is provided (cortisol, 15 to 20 mg in the morning and 5 mg in the evening) until normal function of the hypothalamic-pituitary-adrenal axis is re-established, which usually does not occur until 6 to 24 months after surgery.

In patients with eucortisolism immediately or within a few weeks of surgery, clinical remission is still likely, though CD is more likely to recur than in patients with postoperative hypocortisolism (Fig. 20-17).^45,46

FIGURE 20-17 Analysis of the recurrence of CD in patients in remission after trans-sphenoidal operation. In the upper left, the disease-free survival in the whole group is shown (numbers in brackets represent patients at risk at the end of each year). In the upper right, the disease-free survival according to the presence of a normal (n = 220) or abnormal (n = 239) pituitary CT scan among the 459 patients who were examined before surgery is shown. Patients with an abnormal preoperative CT scan had a higher risk of recurrence, but the difference was of borderline significance (p = 0.053). In the lower left, the disease-free survival according to the early postoperative morning cortisol level is shown for the 482 patients for whom the result was available. Patients with undetectable (n = 94) and low (n = 250) cortisol levels had fewer recurrences during follow-up than did patients with low-normal (n = 97) and high-normal (n = 41) cortisol levels. The overall difference among the four groups was highly significant (p < 0.0001). In the lower right, the disease-free survival according to the length of glucocorticoid substitution therapy is reported for all of the 510 patients who had successful operations. Patients who did not need any substitution therapy (n = 821) and those who required glucocorticoid replacement for less than 1 year (n = 248) showed a higher risk of recurrence than did patients treated for more than 1 year (n = 180). The overall difference among the three groups was highly significant (p < 0.0001).

(Modified from Bochicchio D, Losa M, Buchfelder M. Factors influencing the immediate and late outcome of Cushing’s disease treated by trans-sphenoidal surgery: a retrospective study by the European Cushing’s Disease Survey Group. J Clin Endocrinol Metab. 1995;80:3114-3120.)

Early Repeat Surgery

If results of cortisol measurement shortly after surgery indicate persistent hypercortisolism, immediate repeat surgery should be considered in certain circumstances.³⁹ Early reoperation (within 1 to 6 weeks) can induce remission in the majority of patients, although with a risk of hypopituitarism. Surgical or pathologic confirmation of an ACTH-positive adenoma that is incompletely resected during the initial pituitary exploration is the most significant predictor of success with early re-exploration, in terms of both remission and avoidance of hypopituitarism. Patients who have undergone limited exploration of the pituitary and selective excision of an area that at surgery appears to be, but proves not to have been, a corticotroph adenoma, also are likely to benefit from repeat surgery. On the other hand, patients who undergo an extensive initial exploration and partial resection of the anterior lobe are unlikely to benefit from a repeat operation if there is no ACTH-positive adenoma in the excised specimen.³⁹ With appropriate patient selection early repeat surgery deserves consideration, especially when prompt control of hypercortisolism is required.

Surgical Management of Recurrent Cushing’s Disease

Patients who have remission of hypercortisolism following surgery occasionally develop recurrent CD. What should be done for the patient with recurrent CD after previous surgery? In a report detailing the results of trans-sphenoidal surgery in 31 patients who had previously undergone a trans-sphenoidal operation and two patients who had had previous pituitary irradiation only, in 24 (73%) of the 33 patients, remission of hypercortisolism was achieved by surgery.⁴⁷ Although CT identified an adenoma in only three of the 33 patients, in 20 patients a discrete adenoma was identified at pituitary exploration. The incidence of remission of hypercortisolism was greatest if an adenoma was identified at surgery and the patient received selective adenomectomy (19, or 95% of 20 patients), if there was evidence at surgery or by preoperative imaging that the previous surgical exposure of the pituitary was incomplete (seven, or 78% of nine patients), if an adenoma was seen on preoperative imaging (three of three patients), or if the patient had had prior pituitary irradiation without surgery (two of two patients). In contrast, only five (42%) of 12 patients who received subtotal or total hypophysectomy had remission of hypercortisolism. Surgically induced hypopituitarism occurred in six (50%) of these 12 patients, but in only one (5%) of the 20 patients who underwent selective adenomectomy. Three (13%) of the 24 patients who were in remission from hypercortisolism following repeat surgery developed recurrent hypercortisolism 10 to 47 months postoperatively.

More recent reports have established that the recurrent adenoma is always at, or immediately contiguous to, the site of the adenoma at the original surgery.^48,49 Furthermore, the recurrent tumor usually is invading the dura, usually the cavernous sinus wall contiguous to the former location of the adenoma. In the Dickerman and Oldfield series, at repeat surgery (44 ± 35 months after the initial surgery), in all 43 patients in whom an adenoma had been identified at the initial surgery, the tumor was found at the same site or contiguous to the same site,⁴⁸ indicating that recurrence of CD is from growth of residual cells left in situ at the original surgery and that at repeat exploration the site occupied by the original tumor should be the focus of the exploration. Dural invasion by an ACTH-producing tumor was identified during repeated surgery in 42 (62%) of the 68 patients re-explored after prior surgery and recurrent or persistent CD. In addition, 39 (93%) of these 42 invasive adenomas were located laterally and involved the cavernous sinus. Adenoma invasion of the dura mater was found in 31 (54%) of 57 microadenomas and in all 11 macroadenomas at repeated surgery. The presence of tumor was not detected in 28 of the 59 patients studied with MRI and in none of these 59 patients was dural invasion evident on MRI. The results of this study thus established that recurrent and persistent CD consistently results from residual tumor. At repeated surgery the residual tumor can be found at, or immediately contiguous to, the site at which the tumor was originally found. Unappreciated dural invasion with growth of residual tumor within the cavernous sinus dura, which frequently occurs without residual tumor or dural invasion being evident on MR images or to the surgeon during surgery, is the basis of surgical failure in many patients with CD.

Thus, repeat trans-sphenoidal exploration of the pituitary and treatment limited to selective adenomectomy should be considered in patients with hypercortisolism despite previous pituitary treatment. If an adenoma is identified during surgery, the chance of remission of CD is high and the risk of hypopituitarism is low; however, if no adenoma can be found and partial or complete hypophysectomy is performed, remission of hypercortisolism is less likely and the risk of hypopituitarism is about 50%.

Radiation Therapy

Fractionated radiation of the sella after failed trans-sphenoidal surgery achieves biochemical remission in most patients (80% at 4 years; Fig. 20-18).⁵⁰ Because remission is delayed 6 months to several years after radiation therapy, medical therapy is used to block adrenal⁴⁵ production of glucocorticoids permitting remission of the clinical syndrome of hypercortisolism until the effects of irradiation occur. Hypopituitarism is an expected side effect, but usually occurs 5 to 10 years after treatment. Other risks of sellar irradiation include, in decreasing likelihood, optic neuropathy, oculomotor neuropathy, and secondary neoplasia. Stereotactic radiosurgery may produce an earlier response than fractionated conventional radiation therapy, but whether it will reduce or increase the risk of radiation-induced complications has not been established. Radiosurgery of the sella, or of the focus of known residual tumor, is an option for treatment of CD after unsuccessful surgery, but, as in fractionated therapy remission of hypercortisolism, biochemical remission is delayed 1 to 3 years and the incidence of permanent hypopituitarism is expected to be similar to the incidence with fractionated irradiation therapy, and with new deficiencies being seen as long as 10 years after treatment.^51–53 In patients who have failed microsurgery, gamma knife radiosurgery provides remission in 40% to 80% of patients.^51–53

FIGURE 20-18 Probability of remission of CD in 30 patients treated with pituitary irradiation after unsuccessful transsphenoidal surgery. Response to XRT.

(Modified from Estrada J, Boronat M, Mielgo M, et al. The long-term outcome of pituitary irradiation after unsuccessful transsphenoidal surgery in Cushing’s disease. N Engl J Med. 1997;336:172-177.)

Medical Therapy

“Chemical adrenalectomy,” by medically blocking production of biologically active cortisol by the adrenal cortex, eliminates hypercortisolism and induces remission from the signs and symptoms of CD while awaiting trans-sphenoidal surgery or, more commonly, while awaiting the effects of pituitary irradiation. For this, ketoconazole, which blocks steroidogenesis, is used.^54,55 Its use is limited primarily by hepatic toxicity. Other side effects include reduced androgen production and gynecomastia in men, and a disulfiram reaction. Recent experience using a combination therapy using a new somatostatin analogue, pasireotide, with a dopamine agonist, cabergoline, prior to ketoconozole, proved effective in normalizing urine-free cortisol in 15 of 17 patients with CD.⁵⁶ Yet another option is metyrapone, which inhibits glucocorticoid and mineralocorticoid production, but its clinical usefulness is also limited by toxicity.

Adrenalectomy

Although, adrenalectomy was the treatment for CD in the mid-20th century, the current use of bilateral adrenalectomy is limited to patients in whom other treatments have failed. Adrenalectomy may be used in lieu of irradiation to avoid hypopituitarism in young patients with CD refractory to surgery or it may be combined with irradiation to reduce the risk of development of Nelson’s syndrome, which occurs in 10% to 20% of patients after adrenalectomy for CD.⁵⁷

Nelson’s Syndrome

As the adenomas of CD are under negative feedback, albeit incomplete negative feedback, by high circulating levels of glucocorticoids, plasma ACTH levels increase greatly after adrenalectomy for CD and some ACTH-secreting pituitary tumors aggressively enlarge after correction of hypercortisolism. Nelson’s syndrome, hyperpigmentation (Fig. 20-19) associated with unbridled production of pro-opiomelanocortin and very high levels of plasma ACTH and α-melatonin–stimulating hormone and the appearance of a rapidly growing pituitary macroadenoma after bilateral adrenalectomy for CD, was identified in the era before microneurosurgery, when bilateral adrenalectomy was the treatment of choice for CD. The risk of Nelson’s syndrome after adrenalectomy is 10% to 20% in most series, but is as high as 30% in others, and it may present 1 to 29 years after adrenalectomy.⁵⁸ It is a serious complication that is characterized by uncontrolled tumor growth and parasellar invasion and compression syndromes that are otherwise atypical of CD. The success of surgical treatment Nelson’s syndrome is limited by the large size and invasiveness of these tumors; the incidence of remission from surgery is in the range of 20%,⁵⁹ but is higher if the tumors are detected before parasellar spread. Surgery followed by irradiation is the preferred treatment,⁵⁸ although many of these tumors are refractory to surgical and/or radiation therapy.

FIGURE 20-19 Nelson’s syndrome. Photograph demonstrates the hyperpigmentation associated with excessive ACTH levels, made more noticeable by the patients light eyes and hair (with permission from the patient).