ACTH-independent CS (see Table 17.2): Approximately 10 % of CS is ACTH-independent including adrenal tumors (both benign and malignant) and primary nodular hyperplasia. IPSS is not indicated in cases of ACTH-independent CS where an adrenal etiology is likely [24–26]. Adrenal CS can be easily diagnosed by the combination of low ACTH levels and finding adrenal nodule(s) by CT or MRI.

ACTH-dependent CS (see Tables 17.2 and 17.3): The vast majority of CS is ACTH dependent—largely from two causes: CD (~80 %) and EAS (~10 %) [27]. Many different tumor types both benign and malignant can produce EAS, and many produce other ectopic biomarkers (see Table 17.3) [5, 27–29]. Historically, differentiating CD from EAS had been difficult but that changed when IPSS was introduced [1, 30, 31]. Since then, it has achieved a high level of diagnostic accuracy. This has improved with modifications including bilateral sampling [32, 33], CRH stimulation, and “normalizing” ACTH responses with prolactin measurements. To demonstrate how IPSS is used in clinical practice, a case is presented below.

Clinical suspicion of CS: A 24-year-old woman experienced a 100-lb. weight gain with Cushing’s features, hyperglycemia, hypokalemia, hypertension, and psychoses requiring multiple psychiatric medications (Figs. 17.2 and 17.3). Her history was complicated by a series of undetermined quantities of Depo glucocorticoid injections over the last 2 years for back pain. Despite the history of exogenous glucocorticoids, CS testing was done because of the disproportionality of her findings. Exogenous CS was deemed an unlikely factor based on the rough assessment of dose, frequency, and remote history. In addition, normal cortisol responses to ACTH stimulation testing further diminished this possibility.

Diagnostic CS testing: The diagnostic CS testing showed hypercortisolism based on late-night serum and salivary and 24-h urine cortisol values (see Table 17.4). Loss of the normal glucocorticoid feedback inhibition was found by dexamethasone suppression testing (DST). ACTH independence was excluded by non-suppressed plasma ACTH levels, thus ruling out adrenal CS.

Differential diagnostic testing for CS etiology: Although an adrenal etiology of CS was ruled out by normal plasma ACTH levels (and later incidental abdominal CT imaging), the differentiation between CD and EAS was uncertain because of equivocal pituitary MRI findings and nondiagnostic hepatic mass pathology.

Summary: This case summarizes the strategic use of IPSS in the medical evaluation of CS. Clinical suspicion to screen for CS was based on the rapidity and severity of her symptomatology disproportionate to her exogenous glucocorticoid exposure. The diagnosis of CS was made by multiple screening tests but the etiology was unclear because of equivocal pituitary MRI findings and the discovery of hepatic masses. Adrenal CS was unlikely based on non-suppressed plasma ACTH levels and normal adrenal gland imaging. On transsphenoidal surgery she was found to have a left-sided pituitary adenoma that was resected, following which the cortisol levels returned to normal.

Plasma ACTH: This test will usually diagnose adrenal CS based on a low serum level (see Table 17.2). The combination of a low plasma ACTH with normal but autonomous urinary cortisol occurs in adrenal “incidentalomas.” CD usually has normal or high ACTH levels, whereas EAS has the highest but considerable overlap occurrence.

Adrenal gland CT or MRI: The combination of hypercortisolism with low ACTH should prompt adrenal imaging. The usual finding is adrenal mass(es) usually unilateral but sometimes bilateral. The finding of an adrenal mass in CS without a low ACTH level, however, is not an evidence for an adrenal etiology since these patients may still have CD [44].

High-dose dexamethasone suppression testing (HDDST): CD is based on the premise that most ACTH-secreting pituitary adenomas have an intact but relatively resistant negative feedback loop. In contrast, EAS has a negative feedback loop which is completely resistant to glucocorticoids with the exception of some carcinoid tumors. Unfortunately, using <50 % suppression “cut-point” yields false-positive and false-negative rates between 20 and 30 %, respectively [45].

CRH testing (Ferring Pharmaceuticals, Limhamn, Sweden, under the tradename Acthrel): CRH can be difficult to procure for outpatient testing but it is considered second best for differentiating CD and EAS. The fact that most ACTH-secreting pituitary adenomas (85 %) respond to CRH with an increase in ACTH secretion and most EAS and adrenal CS do not is the basis for this test [46]. Dexamethasone is sometimes used in conjunction with CRH to improve the diagnostic accuracy, but the test has enough false-negative results to limit its use. The major use for this test is diagnosing pseudo-Cushing’s and physiologic hypercortisolemic states (see Table 17.2) [47].

Pituitary MRI (with gadolinium enhancement): Pituitary MRI has a diagnostic accuracy of only about 50 % largely as a result of two problems [48–51]. First, the small size of CD adenomas (sometimes ~1 mm) and second, lack of functional assessment, which is a problem with non-ACTH-secreting pituitary “incidentalomas.” Pituitary incidentalomas have been found in 10–20 % of unselected MRIs and autopsies [52–54]. Co-lateralization of MRI findings with IPSS will further enhance not only diagnostic certainty but also lateralization within the pituitary body [55, 56].

IPSS (90 % diagnostic accuracy): Of all of the etiologic tests for CS, IPSS has the best diagnostic accuracy as reported by high-volume referral centers [57]. The primary use of IPSS is to localize the source of ACTH (see Table 17.7). The secondary and less accurate use of IPSS is to lateralize the side of ACTH source within the pituitary body to allow for more selective and less invasive transsphenoidal surgery. It must be remembered that IPSS is not useful in diagnosing CS. IPSS has proven useful in both pregnancy [58] and young children [17, 59–61]. The diagnostic IPSS criteria to localize and lateralize the source of ACTH are seen in Table 17.8. Despite its successes IPSS still has inaccuracies in otherwise difficult cases [62, 63]. Furthermore, because IPSS requires a high degree of technical skill and expertise, results could vary widely. Anatomic variability precludes about 10 % of CS patients who cannot be successfully catheterized for IPSS. The principal reason for CRH stimulation is to correct for ACTH pulsatility [64]. Pulsatility is also a good reason to draw two baseline ACTH levels. CRH stimulation improves IPSS diagnostic accuracy from 60 to 90 % according to several reports [56, 65, 66]. A systematic analysis that included 21 studies and 569 patients found that IPSS with CRH stimulation achieved 96 % sensitivity and 100 % specificity in discriminating Cushing’s disease from EAS [47]. Interestingly, CRH also stimulates AVP, prolactin, and other ACTH-related peptides [67–69]. CRH has also been combined with TRH (not commercially available) to dually stimulate both ACTH and prolactin [70]. Difficulty procuring CRH has led to the interest in alternatives such as naloxone, metyrapone, and vasopressin [71, 72]. A synthetic version of vasopressin and desmopressin (DDAVP; Ferring Pharmaceuticals) compared to CRH produces less ACTH stimulation [73–75]. Nevertheless, desmopressin stimulation improves diagnostic accuracy over 90 % and comparable to CRH [76]. Combined CRH and AVP stimulation also yields excellent diagnostic accuracy and perhaps better than either alone [77, 78]. The desmopressin technique is similar to CRH and 10 mg of desmopressin is used in place of the CRH. IPSS prolactin levels can be used to correct for venous effluent concentrations as they lateralize in the majority of non-CS subjects showing the need to correct for difference in venous effluent [79–84]. When petrosal-to-peripheral ACTH ratios are “normalized” to prolactin petrosal-to-peripheral ratios, the diagnostic accuracy for both localization and lateralization improves [82, 85]. “Normalization” of ACTH gradients improves accuracy not only prospectively but retrospectively using archived IPSS blood samples [79, 84]. In the current case, ACTH gradients lateralized to the left (Table 17.6) only after prolactin normalization, which is where the CD adenoma was found. Beyond prolactin, other IPSS neuroendocrine biomarkers and pituitary hormones have yet to become useful [69, 86, 87]. IPSS lateralization refers to determining the side, either right or left, of the CD adenoma within the pituitary body. Lateralization unfortunately has not achieved the diagnostic accuracy of localizing the source of ACTH, either within the pituitary body or not. Some studies have shown that a gradient of ≥1.4 between the ACTH concentrations in the two sinuses predicted the side of the tumor with up to 71 % accuracy if catheters were appropriately placed [48, 49, 88–92]. However, others have not found such strong predictive values. Since there is a 50 % chance of correctly predicting the location without the aid of any anatomical data, this expensive procedure cannot be justified for lateralization alone [93–95]. If MRI lateralizes a pituitary lesion to the same side, the accuracy is much higher when it does not. If the IPPS lateralizes but the MRI does not, the whole pituitary body should still be explored at TSS since ~30 % have contralateral tumors [91]. The reasons for failing to lateralize are listed in Table 17.9. Besides the usual considerations of variations in venous drainage and catheter placement, an intrinsic lateralization of pituitary function favors the right of the time 80 %. This may explain why left-sided IPSS lateralization is most accurate for CD. Other causes are seen in the table and should be considered when evaluating the data.