Pitfalls in Performing and Interpreting Inferior Petrosal Sinus Sampling: Personal Experience and Literature Review

Abstract

Context

Inferior petrosal sinus sampling (IPSS) helps differentiate the source of ACTH-dependent hypercortisolism in patients with inconclusive biochemical testing and imaging, and is considered the gold standard for distinguishing Cushing disease (CD) from ectopic ACTH syndrome. We present a comprehensive approach to interpreting IPSS results by examining several real cases.

Evidence Acquisition

We performed a comprehensive review of the IPSS literature using PubMed since IPSS was first described in 1977.

Evidence Synthesis

IPSS cannot be used to confirm the diagnosis of ACTH-dependent Cushing syndrome (CS). It is essential to establish ACTH-dependent hypercortisolism before the procedure. IPSS must be performed by an experienced interventional or neuroradiologist because successful sinus cannulation relies on operator experience. In patients with suspected cyclical CS, it is important to demonstrate the presence of hypercortisolism before IPSS. Concurrent measurement of IPS prolactin levels is useful to confirm adequate IPS venous efflux. This is essential in patients who lack an IPS-to-peripheral (IPS:P) ACTH gradient, suggesting an ectopic source. The prolactin-adjusted IPS:P ACTH ratio can improve differentiation between CD and ectopic ACTH syndrome when there is a lack of proper IPS venous efflux. In patients who have unilateral successful IPS cannulation, a contralateral source cannot be excluded. The value of the intersinus ACTH ratio to predict tumor lateralization may be improved using a prolactin-adjusted ACTH ratio, but this requires further evaluation.

Conclusion

A stepwise approach in performing and interpreting IPSS will provide clinicians with the best information from this important but delicate procedure.

Cushing syndrome (CS) can be divided into ACTH-dependent and ACTH-independent hypercortisolism. Inferior petrosal sinus sampling (IPSS) is the “gold standard” to determine the source of hypercortisolism in ACTH-dependent CS. IPSS is indicated following inconclusive imaging but positive, biochemical testing confirming ACTH-dependent hypercortisolemia (1, 2). The sensitivity and specificity of IPSS in diagnosing Cushing disease (CD) can range from 88% to 100% and 67% to 100%, respectively (3). IPSS may not be required in patients who have a sellar mass > 6 mm when dynamic testing is suggestive of CD. A recent study by Yogi-Morren et al. that included 104 patients with CD and 26 patients with ectopic ACTH syndrome (EAS) showed that pituitary tumors > 6 mm have a 96% specificity for CD (vs. EAS) (4, 5). Many clinicians have limited experience performing and interpreting IPSS results. We provide readers a summary of the current literature and offer strategies to help optimize use of this important diagnostic test.

History of IPSS

IPSS was first performed at the Walter Reed Army Medical Center in 1977, with selective and bilateral (but not simultaneous) catheterization via a transjugular approach (6). This evolved to a multistep process that included unilateral sampling (for localization) followed by bilateral and simultaneous catheterization (for localization and lateralization) (7). In 1985, Oldfield and colleagues demonstrated a high rate of remission following hemihypophysectomy to the side showing highest ACTH concentrations on IPSS despite lack of radiological evidence of microadenoma, although the number of patients was small (8). The utility of IPSS has since been further improved by direct stimulation of the corticotroph adenoma using corticotropin-releasing hormone (CRH) or vasopressin (9, 10). In 1991, Oldfield and his National Institutes of Health (NIH) colleagues published on 281 patients with CS, a report that supported the use of CRH to distinguish CD from EAS (11). Prolactin measurement was suggested to measure pituitary venous effluent and confirm successful cannulation in 2004 (12).

IPSS Procedure

Inferior petrosal sinus sampling is most often performed under local anesthesia. The femoral veins are cannulated, and microcatheters are placed in the bilateral petrosal sinuses under fluoroscopic guidance. Venous angiography is used to confirm correct catheter placement, as demonstrated by retrograde flow of contrast into the contralateral cavernous sinus (3). Simultaneous venous blood samples are obtained from the petrosal sinuses and a peripheral vein. The blood samples are drawn at baseline and then at 2, 5, 10, and 15 minutes following CRH (1 μg/kg, maximum 100 μg) or vasopressin (10 μg) administration. Some institutions may use slightly different timepoints for collection.

IPS catheter placement is reassessed at various intervals during the procedure to ensure sustained cannulation. Our neuroradiologist checks the position of the IPS catheters at 5 minutes and once sampling is completed. If the catheter locations appear unstable, imaging is obtained prior to each blood draw. The patient is observed for several hours postprocedure. There are rare reports of serious complications, the most notable being stroke and subarachnoid hemorrhage (13-16). These were believed related to aberrant anatomy, leading to transient hypotension and/or vascular injury during the procedure (15). IPSS should be aborted for any significant intraoperative hypotension, hypertension, or change in neurological status. The most common adverse events are groin hematoma (4%) and transient headache (occurs more often if the sinus is small) (17). IPSS is a technically challenging procedure and should be performed by an experienced interventional or neuroradiologist. Although prospective, randomized data are missing, we and others expect better results from centers that have a dedicated radiologist who focuses on successful IPSS. A stepwise approach to IPSS interpretation is given in Fig. 1.

Figure 1.

Our approach to differentiating the etiologies of endogenous, ACTH-dependent Cushing syndrome.

Some centers use DDAVP (desmopressin) for IPSS because of lower cost and past unavailability of CRH. DDAVP stimulates ACTH production via the V2R receptor subtype on corticotroph adenomas and elicits a response similar to CRH (18-20). There is a theoretical risk of false-positive results (positive IPS-to-peripheral [IPS:P] ACTH gradient) using DDAVP in a patient with an ectopic, ACTH-producing tumor if the V2R receptor is present (21). However, the clinical experience seems similar using both CRH and DDAVP. Deipolyi et al. showed comparable diagnostic results using identical IPS:P ACTH gradients after DDAVP and CRH stimulation (10). An earlier study by Tsagarakis et al. demonstrated good IPSS sensitivity and specificity in 54 patients following administration of combination CRH and DDAVP (22).

It is our preference to perform IPSS before starting any antihypercortisolemic therapy, but patients may be referred to our center after already having been prescribed these medications. If we determine that a patient on antihypercortisolemic therapy needs IPSS, our practice is to stop the medication and schedule the procedure once we confirm hypercortisolemia on biochemical testing.

Anatomical Variations

There are anatomical variations in pituitary drainage that can affect the success and interpretation of IPSS. The anterior pituitary drains from the cavernous into the inferior petrosal sinuses. The inferior petrosal sinuses course posterior and caudal and enter the jugular veins at the skull base (23) (Fig. 2). It is not uncommon to have unequal drainage of the cavernous sinuses. Some literature suggests this anatomical variant may be present in 40% of individuals (24, 25). There are other aberrations that make analysis much more difficult. The most notable variants are as follows: (1) inferior petrosal anastomosis to vertebral venous plexus; (2) no connection between the inferior petrosal sinus and the jugular vein; and (3) a hypoplastic inferior petrosal sinus (26). Previous studies have suggested that microcatheters can be used to obtain accurate venous sampling if abnormal anatomy is known or suspected (27). It is also possible to obtain a direct sample from the cavernous sinuses, but this approach has not proved superior to IPSS (28-32). The literature indicates that the prevalence of IPS abnormalities exceeds that of false-negative IPSS results, meaning that adequate catheterization is most often feasible.

Figure 2.

In this schematic, the infundibulum and pituitary gland are marked in red. The pituitary gland sits in the sella. The relevant venous structures are designated in blue.

Case Examples

Case Presentation 1

A 50-year-old woman with progressive weight gain, malaise, and arthralgias presented for evaluation. Her medical history was notable for uncontrolled type 2 diabetes mellitus, steatohepatitis, hypertension, hyperlipidemia, and obstructive sleep apnea. Physical examination showed abdominal obesity, facial plethora, and diffuse acanthosis nigricans. Biochemical testing revealed 24-hour urinary free cortisol (UFC) levels of 94.2 and 152.2 µg (0-50), ACTH levels of 50 and 69 pg/mL (normal range, 10-60 pg/mL), and a cortisol level of 13.8 µg/dL after 1-mg dexamethasone suppression test (DST). Sellar magnetic resonance (MR) imaging demonstrated a 3 × 4-mm focus of early enhancement on the left side of the gland. IPSS was performed using DDAVP; the neuroradiologist reported successful bilateral IPS catheter placement (Table 1).

Table 1.

IPSS results for case presentation 1

Time (min)	ACTH
	Peripheral (P)	Petrosal sinus
		Right	Left	ACTH ratio
				R/P	L/P
Baseline	52	811	469	15.6	9.0
+3	2	3015	271	1507.5	135.5
+5	59	4558	741	77.3	12.6
+10	90	2359	414	26.2	4.6
+15	90	1696	596	18.8	6.6

Abbreviations: L/P, left IPS/peripheral; R/P, right IPS/peripheral.

Can IPSS differentiate between ACTH-dependent and pseudo-Cushing syndrome?

No, IPSS cannot differentiate between CD and pseudo-Cushing states. A diagnosis of ACTH-dependent CS must be established before a patient is referred for IPSS.

It is critical to establish a diagnosis of endogenous hypercortisolism before pursuing IPSS. This is an invasive procedure and the presence of a positive IPS:P ACTH ratio may precipitate unnecessary surgical exploration. Pseudo-Cushing states have a number of clinical associations including: pregnancy, morbid obesity, severe psychological stress (including major depressive disorder), uncontrolled diabetes mellitus, chronic alcoholism, physical illness or injury, and severe sleep apnea (33). It is thought that higher brain centers stimulate CRH secretion in these conditions, leading to activation of the entire hypothalamic–pituitary–adrenal axis (34). The negative feedback inhibition of cortisol on CRH and ACTH may restrain the resultant hypercortisolemia. It is uncommon to see UFC elevations > 4 times the upper limit of normal in patients with pseudo-Cushing states (1). These same conditions can blunt the circadian rhythm and cause abnormal late-night salivary cortisol concentrations (35-38). Repeat biochemical testing following resolution or treatment of the underlying disease process should be considered. Additional confirmatory testing such as 2-day low-dose DST or a combined dexamethasone suppression/CRH stimulation test may also be used to exclude pseudo-Cushing states. The endocrinology team felt the hypercortisolism in this patient was not secondary to a pseudo-Cushing state and proceeded with an IPSS study.

Researchers at the NIH have demonstrated that there is significant overlap of ACTH concentrations and IPS:P ACTH ratios between CD, normal volunteers, and pseudo-Cushing states (39, 40). Yanovski et al. [37] compared IPSS results from patients with CD (n = 40) to normal subjects (n = 7) and those with pseudo-Cushing states (n = 8) at baseline and 2 minutes post-CRH stimulation (the institutional review board did not allow additional sampling). There was crossover in IPS:P ACTH ratios between the various conditions. This study confirms that IPSS should not be used to distinguish pseudo-Cushing states from CD. Pseudo-Cushing states must be eliminated as potential causes of ACTH-dependent hypercortisolism before referring a patient for IPSS (40).

Is the IPSS result in this patient consistent with Cushing disease?

Yes, the patient has a central to peripheral ACTH ratio > 3 following CRH stimulation. This result confirms CD. She should be referred to an experienced neurosurgeon for surgical exploration of the pituitary gland.

Cushing disease is confirmed when the basal IPS:P ACTH ratio is > 2 and/or the CRH-stimulated ratio is > 3 (8, 41-43). The peripheral ACTH level at 3 minutes post-CRH administration was low (2 pg/mL) and consistent with laboratory error considering her other ACTH values. ACTH must be collected in an EDTA tube and transported swiftly to the laboratory on ice. The sample should then be centrifuged to separate plasma and frozen (if not processed). Delayed or improper handling of the samples during transportation may result in falsely low levels (44). In theory, if a pituitary tumor lacked any CRH receptors, one may not see a rise in ACTH following CRH stimulation. In such an event, the IPS:P ACTH ratio may still be elevated before CRH administration. We have not observed such a clinical scenario in the past (positive IPS:P ACTH ratio before but not after CRH stimulation). This might be because even a low density of CRH receptors on corticotroph cells is enough to generate an increased gradient after CRH stimulation in the highly concentrated IPS environment.”

The patient had successful transsphenoidal resection of a left-sided pituitary mass. Her cortisol nadired following the procedure at 1.7 µg/dL and she required postoperative glucocorticoids. An ACTH-positive corticotroph adenoma was confirmed by surgical pathology.

Is it necessary to confirm hypercortisolism at the time of IPSS?

Yes, IPSS may not be reliable if there is lack of hypercortisolemia at the time of the procedure (39, 45-47).

The absence of sustained hypercortisolism can cause misleading IPSS results. This can happen in a patient with cyclical CD if cortisol is tested during a trough period. The IPSS may reveal an absent IPS:P ACTH gradient because of lack of ACTH production by the corticotroph adenoma and suppression of normal corticotrophs secondary to recent hypercortisolism. This IPSS result may mimic an ectopic source. By contrast, incomplete suppression of normal corticotroph function caused by intermittent ACTH production from a cyclical ectopic tumor might produce a false-positive IPS:P ACTH ratio (48, 49). For these reasons, we measure serum cortisol the morning of scheduled IPSS and proceed only if the value is > 10 μg/dL (50). There are some centers that use midnight salivary cortisol to confirm hypercortisolism the night before IPSS (45, 46). We know of several investigators who require 6 weeks of consistent hypercortisolemia via 24-hour UFC before performing the procedure. This approach requires close monitoring of patients and clinical significance, cost, and practicability require more detailed study.

Table 2 gives an example of IPSS results obtained in a second patient with cyclical Cushing tested during a trough period (49). The patient was confirmed to have cyclic CD by serial measurements of 24-hour UFC, demonstrating 3 elevated and 2 normal/low values (Table 3). IPSS was performed using CRH. There was unsuccessful left-sided IPS cannulation (Table 2). Despite the significant IPS:P gradient on the right side, the authors felt the results were uninterpretable because of a lack of hypercortisolism during the procedure. This is evidenced by the peripheral cortisol and ACTH levels of 6.2 μg/dL and 8 pg/mL at the time of sampling, respectively (49). Swearingen et al. reported 3 cases of false-positive IPSS findings (IPS:P ACTH gradient suggestive of CD) in ectopic tumors. These patients had low peripheral ACTH levels obtained during IPSS, making the IPS:P ratios difficult to interpret (51). After 2 IPSS attempts, during which he was not hypercortisolemic, our patient was referred for pituitary exploration. The transsphenoidal surgery (performed through the nose) yielded an ectopic corticotroph adenoma suspended from the mucosa of the sphenoid sinus (49). The IPSS results are difficult to interpret even retrospectively because we cannot know the venous drainage pattern of this ectopic adenoma. In these rare situations, IPSS may be consistent with either CD or an ectopic ACTH syndrome.

Table 2.

IPSS results obtained during a trough period in our patient with cyclical Cushing

Time (min)	ACTH (pg/mL)					Cortisol (μg/dL)
	Peripheral (P)	Petrosal sinus				Peripheral (P)
		Right	Left	ACTH ratio
				R/P	L/P
-10	8	73	9	9.1	1.1	6.2
-5	8	64	8	8	1
+2	7	145	7	21	1
+5	7	386	7	55	1
+10	7	471	8	67	1.1
+15	9	2	10	0.2	1.1

Abbreviations: L/P, left IPS/peripheral; R/P, right IPS/peripheral.

Table 3.

Biochemical testing for our patient diagnosed with cyclical ectopic corticotroph adenoma (49)

Measurements (units)	Time
	Presentation	Day 2	Week 5	Week 17	Week 18	Week 19	Week 22
Urinary free cortisol (mcg/d)ª	119.6 (RAI)	110.2 (RAI)	35.9 (RAI)	343.0 (RAI)	789.1 (HPLC)	7.3 (HPLC)	692.4 (RAI)
Urinary creatinine (g/d)	1.7	1.6	1.6	1.9	1.9	1.8	1.8
Urinary volume (mL)	1329	1318	1382	1900	1250	2200	1300

RAI normal range: 20-100 mcg/d. HPLC normal range: < 45 mcg/d.

Abbreviation: RAI, radioactive iodine.

^aUrinary free cortisol measurements were obtained by either RAI or HPLC.

Case Presentation 2

A 51-year-old woman with progressive weight gain, hair loss, emotional lability, and new-onset type 2 diabetes mellitus was referred for possible hypercortisolism. Physical examination revealed abdominal obesity, widespread ecchymoses, excess supraclavicular fat, mild muscle wasting, and proximal muscle weakness. Her biochemical results were as follows: 24-hour UFC, 2000 µg (0-50); morning cortisol 43.8 µg/dL; and ACTH 170 pg/mL following 8 mg DST. Sellar MR imaging (MRI) showed a 5-mm right-sided adenoma. She had an unremarkable chest computed tomography (CT) scan. IPSS was performed using CRH; the neuroradiologist reported successful bilateral IPS catheter placement (Table 4).

Table 4.

IPSS results for case presentation 2

	ACTH (pg/dL)					Prolactin (ng/mL)					Cortisol (μg/dL)
	Peripheral (P)	Petrosal sinus				Peripheral (P)	Petrosal sinus				Peripheral (P)
		RT	LT	ACTH ratio			RT	LT	PRL ratio
Time (min)				R/P	L/P				R/P	L/P
-10	102	125	124	1.2	1.2	4.8	5.7	8.4	1.2	1.7	64
-5	100	222	111	2.2	1.1	5.6	8.3	6.3	1.5	1.1
+2	102	130	129	1.3	1.2	5.0	6.2	7.1	1.2	1.4
+5	143	220	191	1.5	1.3	5.3	6.3	6.7	1.1	1.2
+10	176	217	235	1.2	1.3	5.2	5.2	6.6	1.0	1.2
+15	197	224	232	1.1	1.2	5.0	5.5	8.5	1.1	1.1

Abbreviations: LT, left; L/P, left IPS/peripheral; RT, right; R/P, right IPS/peripheral.

Is the lack of an IPS:P ACTH gradient consistent with a diagnosis of ectopic CS in this patient?

Absent a significant IPS:P ACTH ratio, the IPS prolactin level should be used as a surrogate marker of appropriate catheterization/normal IPS venous efflux to avoid a false-negative result.

IPS prolactin can help confirm correct catheter placement during venous sampling. IPS catheterization can be verified using an IPS:P prolactin ratio ≥ 1.3 to 1.8 before and after CRH/DDAVP administration (12, 52). The presence of an IPS:P prolactin ratio < 1.3 and the absence of significant elevation should raise suspicion for either improper cannulation or abnormal IPS venous efflux (53). Prolactin is the most abundant anterior pituitary hormone and spatial separation of normal lactotrophs from corticotropes within the gland make it a good reference hormone (54). Previous studies have advocated use of GH, B-endorphin, and TSH for localization of ACTH-secreting microadenomas, but interpreting these data remain challenging (55). These hormones may also be suppressed in hypercortisolism (most often ectopic ACTH syndrome) (56-58).

There are several possible reasons that radiological confirmation of IPS cannulation and IPS:P prolactin ratios might be discordant. This incongruity is more often attributable to procedural abnormalities rather than sensitivity of the IPS:P prolactin ratio (52). The most common procedural problems include: (1) intermittent displacement of the catheter tip (from the IPS); (2) aspiration of samples from an aberrant collateral vessel (despite proper cannula placement); and (3) the sample does not actually contain venous blood from the pituitary gland (occurs when too small a catheter is used to interrupt laminar flow). As with other technically demanding procedures, the success of IPSS often depends upon the experience of the interventional or neuroradiologist (59).

The absence of an IPS:P ACTH gradient here (Table 4) suggests an ectopic source of hypercortisolism. The only exception is the right IPS:P ACTH ratio at -5 minutes (2.2), a borderline positive value for CD. The neuroradiologist confirmed catheter placement using fluoroscopy. We routinely check IPS prolactin levels at our institution to ensure proper IPS venous efflux before concluding that the ACTH source is ectopic. Some centers limit procedural costs by storing serum during IPSS and checking prolactin concentrations only if ACTH gradients are absent.

What is the role of the prolactin-adjusted IPS:P ACTH ratio when there is lack of IPS catheterization or appropriate IPS venous efflux?

The prolactin-adjusted IPS:P ACTH ratio can improve differentiation between CD and ectopic ACTH syndrome in the absence of proper IPS venous efflux.

Findling and colleagues used prolactin as an index of pituitary venous efflux in 3 cases of surgically proven CD in whom IPSS failed to demonstrate an appropriate IPS:P ACTH gradient (and an ectopic source could not be found). The authors showed that a prolactin-adjusted IPS:P ACTH ratio (dominant post-CRH IPS:P ACTH/ipsilateral pre-CRH IPS:P prolactin) > 0.8 would have identified these 3 patients as having CD (12). The ratio was < 0.6 in 5 EAS patients. A 2013 retrospective study from the NIH examined prolactin-adjusted IPS:P ACTH ratios in CRH-stimulated IPSS samples from 29 patients with ACTH-dependent CS (60). They diagnosed CD using a prolactin-adjusted IPS:P ACTH ratio ≥ 1.3. All ratios ≤ 0.8 corresponded to ectopic ACTH syndrome. Prolactin-adjusted IPS:P ACTH ratios ranging from 0.8 to 1.3 did not discriminate between CD and ectopic ACTH syndrome. The authors concluded that a prolactin-adjusted IPS:P ACTH ratio of: (1) ≤ 0.8 suggests EAS; (2) ≥ 1.3 indicates CD; and (3) 0.8 to 1.3 needs further investigation (60). These findings are similar to those of Grant et al (61).

The use of the prolactin-adjusted ACTH ratio has potential limitations. The pre-CRH prolactin value does not account for erroneous IPS sampling that may occur after CRH injection because of repositioning of the catheter tip. It is also possible that the prolactin-adjusted IPS:P ACTH ratio in ectopic ACTH syndrome may mimic CD if there is unsuccessful IPS cannulation (52). In case 2, using the baseline sample at -10 minutes, the prolactin-adjusted IPS:P ratio is 0.77 and below the threshold of 0.8, which could indicate an ectopic source. If we use the second pre-CRH prolactin value (drawn at the -5 minute timepoint), then the ratio further decreases to 0.63 (very close to the 0.6 cutoff suggested by Findling et al.) (52). Larger studies are needed to confirm the role of the prolactin-adjusted IPS:P ACTH ratio in ACTH-dependent CS and failed IPS cannulation.

Are there additional clues that may suggest lack of proper IPS venous efflux during IPSS?

Yes, the absolute ACTH levels (pre- and post-CRH) in the inferior petrosal sinuses and ACTH response to CRH in the periphery may identify CD when there is lack of a significant central ACTH gradient.

Wind et al. showed that baseline ACTH values < 200 pg/mL and peak post-CRH ACTH values < 400 pg/mL may indicate lack of proper IPS venous efflux (62). The index patient only achieved an ACTH > 200 pg/mL (in the right IPS) at -5 minutes, corresponding to a positive IPS:P ACTH ratio (> 2) at the same timepoint. Additional data suggest that peripheral ACTH response to CRH may help identify false-negative results (the presence of CD but absence of a central IPS:P ACTH gradient) (46, 63). Swearingen et al. studied 9 patients with negative IPSS results and no known ectopic source and evaluated peripheral ACTH response to CRH. Eight of these patients had a significant rise in peripheral ACTH (> 35%) following CRH administration. All 9 patients were later proven to have CD. These findings support using CRH-stimulated peripheral ACTH levels to improve the diagnostic accuracy of IPSS (51). In a recent study from Germany, an increase of ≥ 43% in peripheral ACTH after CRH administration had an 83% sensitivity and 94% specificity for CD (64). Previous studies that compared standard CRH-stimulation testing (measuring cortisol) to IPSS yielded mixed results (65, 66). In our case, the patient had a significant increase in peripheral ACTH following CRH administration, signalling a probable false-negative IPSS result (Table 4).

Our patient elected to pursue transsphenoidal resection of the right-sided pituitary lesion. She was aware that it might have been an incidental finding. Her cortisol and ACTH levels decreased to normal (11.4 μg/dL and 19 pg/mL) during the immediate postoperative period but she did not become hypocortisolemic. Pathology confirmed the presence of a corticotroph adenoma. The patient experienced mild symptoms of (relative) adrenal insufficiency and was discharged on physiologic hydrocortisone. She decided to see her local endocrinologist for further postoperative care and surveillance.

Case Presentation 3

A 30-year-old woman with progressive weight gain, fatigue, muscle weakness, and skin changes suggestive of hypercortisolism presented for endocrine evaluation. Her biochemical testing was as follows: 24-hour UFCs of 824 and 3515 µg (0-50); random plasma ACTHs of 80 and 102 pg/mL, and cortisol of 44.8 µg/dL after 1 mg DST administration. Sellar MR failed to reveal a pituitary adenoma. This patient’s whole-body imaging did not show any source of possible ectopic ACTH syndrome. She underwent IPSS using CRH (Table 5), during which the neuroradiologist reported successful bilateral IPS catheter placement.

Table 5.

IPSS results for case presentation 3

Time (min)	ACTH (pg/mL)					Prolactin (ng/mL)					Cortisol (μg/dL)
	Peripheral (P)	Petrosal sinus				Peripheral (P)	Petrosal sinus				Peripheral (P)
		RT	LT	ACTH ratio			RT	LT	PRL ratio
				R/P	L/P				R/P	L/P
-10	73	79	89	1.1	1.2	14	13.4	24.1	1.0	1.8	29.1
-5	76	77	85	1.0	1.1	13.1	12.5	28.6	1.0	2.2
+2	83	108	117	1.3	1.4	13.3	12.5	35.2	1.1	2.8
+5	210	196	267	0.93	1.3	13.6	11.8	38.8	1.2	3.3
+10	246	326	315	1.3	1.3	13	13.3	34.6	0.98	2.6
+15	288	383	370	1.3	1.3	13.7	12.7	34.9	1.1	2.7

Abbreviations: LT, left; L/P, left IPS/peripheral; RT, right; R/P, right IPS/peripheral.

Is this IPSS result consistent with ectopic ACTH syndrome?

No; the presence of a pituitary adenoma on the contralateral side cannot be excluded when there is only unilateral successful IPS catheterization.

The IPS:P prolactin ratios in this patient indicate lack of appropriate venous efflux or failed cannulation on the right side (all < 1.3). Her IPSS results demonstrate, however, significant peripheral ACTH response to CRH, suggestive of central disease. She had transsphenoidal exploration, which revealed a right-sided adenoma. The tumor stained positive for ACTH on surgical pathology. The patient became hypocortisolemic during the immediate postoperative period confirming CD (2). If there are no intersinus anastomoses, there may not be a positive contralateral IPS:P ACTH gradient in CD (when there is no mixing of blood, the ACTH values represent suppressed normal corticotrophs). Our neuroradiologist has encountered this only once in his 15 years of experience. In the absence of a positive IPS:P ACTH ratio, a unilateral successful IPS cannulation cannot rule out the presence of a corticotroph adenoma on the contralateral side.

Case Presentation 4

A 56-year-old woman with progressive weight gain, easy bruising, and fatigue was evaluated for possible CS. Physical examination revealed abdominal obesity, wide and violaceous striae, excess dorsocervical and supraclavicular fat, facial plethora, and thin extremities. Her medical history was notable for type 2 diabetes mellitus, hypertension, and hyperlipidemia. Her biochemical results showed midnight salivary cortisols of 203 and 341 ng/dL (normal, < 90); 24-hour UFC, 27.8 µg (0-50); morning cortisol, 27.2 µg/dL; and morning ACTH, 126 pg/mL. Sellar MR demonstrated bilateral adenomas (right = 4 mm, left = 2 mm). CT scan of the chest and adrenals was negative. The patient was referred for IPSS. The procedure was performed using CRH and the neuroradiologist reported successful bilateral IPS catheter placement (Table 6).

Table 6.

IPSS results for case presentation 4

Time	ACTH (pg/mL)			ACTH ratio	PRL (ng/mL)			ACTH/PRL ratio		PRL-adjusted ACTH ratio
	Peripheral	RT	LT	LT/RT	Peripheral	RT	LT	RT	LT	RT/LT
-10	77	1192	3295	2.8	17.5	17.5	435.0	68.1	7.6	9.0
-5	65	1585	2710	1.7	16.4	43.4	457.5	36.5	5.9	6.2
+2	81	2744	12 806	4.7	15.6	21.9	282.0	125.3	45.4	2.8
+5	131	5526	14 230	2.6	16.3	35.8	391.5	154.4	36.4	4.3
+10	153	5870	9058	1.5	15.6	30.4	389.0	193.1	23.3	8.3
+15	101	4717	10 840	2.3	14.5	27.0	366.5	174.7	29.6	5.9

Abbreviations: LT, left; PRL, prolactin; RT, right.

What is the value of the intersinus ACTH ratio in tumor lateralization?

An intersinus ACTH gradient > 1.4 has limited value in predicting tumor lateralization (8).

There are limited data on the use of IPSS to lateralize pituitary adenomas. In 1985, Oldfield and colleagues proposed an intersinus ACTH gradient of > 1.4 (before CRH administration) for ipsilateral lateralization of pituitary tumors (8). This ratio has proven less reliable as more data have emerged, including a large series of 501 patients published by NIH in 2013 (62). The literature reports correct tumor lateralization using this ratio in 50% to 70% of cases (62, 67-69). There is no evidence to suggest that lateralization improves following CRH administration. In later work from the NIH by Wind et al., the positive predictive value (PPV) for accurate lateralization peaked at 86% using an intersinus ACTH gradient of 60:1. The authors acknowledged, however, a substantive limitation because only 7% of patients who had a unilateral adenoma reached this ratio. This same paper showed that left-sided ACTH lateralization was associated with greater accuracy (reasons unknown). The highest level of accuracy was observed in patients who had consistent lateralization before and after CRH administration (PPV = 72%) (62). These studies reinforce the need for careful neurosurgical exploration of the pituitary gland to identify an adenoma, which may be smaller than 1 mm in the largest diameter (70). This patient’s intersinus ACTH ratio suggested a left-sided lesion. This was inaccurate. Transsphenoidal exploration revealed a right-sided adenoma. The surgical pathology confirmed a corticotroph adenoma.

Does the prolactin-adjusted intersinus ACTH gradient improve lateralization?

The use of a prolactin-adjusted intersinus ACTH gradient > 1.4 improves corticotroph adenoma lateralization during IPSS.

There have been several attempts to improve the predictive value of the intersinus ACTH gradient. In 2012, Mulligan et al. (59) showed improved adenoma lateralization from 54% to 75% using a prolactin-adjusted intersinus ACTH gradient of > 1.4 in their series. The combination of data from pituitary MRI and prolactin-adjusted intersinus ACTH ratio enhanced the lateralization concordance to 82% (55). When successful bilateral IPS catheterization was confirmed using an IPS:P ACTH ratio > 1.3 (n = 14), there were no instances in which the prolactin-adjusted IPS:P ACTH ratio was associated with a contralateral tumor (adenoma was either ipsilateral or centrally located) (59). A later analysis by Qiao et al. again supported improved tumor lateralization using a prolactin-adjusted intersinus ACTH gradient (53). They increased their lateralization from 65% to 77% using an intersinus prolactin-adjusted ACTH > 1.4 following DDAVP administration. They suggested that anatomical variation (such as preexisting communication between the cavernous sinuses) might explain sampling failures (53). They did not comment on the ability of the intersinus prolactin-adjusted ACTH ratio to predict lateralization in patients who had successful bilateral IPS cannulation based on the IPS:P prolactin ratio. There are certain (apparent) situations in which this correction cannot be applied, including rare cases of corticotroph hyperplasia and ectopic pituitary tumors (59). In our patient, the intersinus prolactin-adjusted ACTH ratio correctly indicated a right-sided adenoma.

There are 2 recent studies that are less supportive of using the prolactin-adjusted intersinus ACTH gradient for tumor lateralization but both are small. De Sousa et al. published a retrospective review of IPSS lateralization in 13 patients (71). Their predicted and surgical findings were concordant in only 4 patients regardless of whether the intersinus gradient was corrected for prolactin. The authors concluded that the adjusted ratio could not be used because of consistent co-lateralization of prolactin and ACTH. The study however successfully cannulated both IPSs in only 7 patients (54%). They did not report separately on patients who had adequate sampling and pathology-proven CD (71). A second paper (72) included only 8 patients, 5 of whom had an adenoma > 6 mm (for whom IPSS may not have been indicated). Of the 8 patients, only 1 demonstrated discordant intersinus ACTH and prolactin-adjusted intersinus ACTH ratios. In summary, further evaluation of the prolactin-adjusted ACTH ratio is needed to prove reliable surgical guidance. A careful surgical exploration of the entire pituitary gland in cases where sellar imaging does not reveal a distinct adenoma is needed.

Special Considerations

IPSS in CRH-producing ectopic tumors

CS resulting from ectopic CRH production is very rare. It can be seen in neuroendocrine tumors arising from the pancreas, adrenal glands, or lungs (47, 73-76). There are 2 case reports of IPSS performed on a combined CRH/ACTH-producing pheochromocytoma and a CRH-producing bronchial carcinoid tumor (47, 76). The patient with a pheochromocytoma had a basal IPS:P ACTH gradient < 2 but a peak IPS:P ACTH gradient of 9.1 following CRH administration. The authors theorized that ectopic CRH production prevented suppression of normal corticotrophs that led to a false-positive result (76). The patient with a bronchial carcinoid tumor had a basal IPS:P ACTH gradient > 2, indicative of central disease (the authors did not use CRH or DDAVP). There were no abnormalities on the sellar CT scan and sellar MR was not performed. Surgical exploration of the pituitary gland failed to demonstrate an adenoma. Histopathological examination revealed pituitary hyperplasia. The ectopic CRH production was believed to prevent normal corticotroph suppression leading to a false-positive IPSS result (47).

Inadvertent use of ACTH instead of CRH during IPSS

There have been reports of inadvertent substitution of ACTH for CRH during IPSS. The trade name for ovine CRH (ACTHrel) is similar to ACTH (commercial name Cosyntropin). Carroll et al. analyzed 3 separate cases of IPSS results following accidental cosyntropin administration (77). All patients demonstrated a decrease in IPS ACTH levels after receiving synthetic ACTH. The authors hypothesized that this might be explained by the pharmacokinetics of the immunometric assay used to measure ACTH (78). An IPS:P ACTH ratio that falls following CRH administration should alert clinicians to possible accidental use of cosyntropin. A separate check to confirm medications to be used followed by a time out immediately before CRH administration (similar to that used in the operating room) may eliminate this issue (79).

The role of jugular venous sampling in ACTH-dependent CS

Internal jugular vein sampling (JVS) has been proposed as an easier and safer alternative to IPSS (80-83). Unfortunately, JVS is less sensitive for diagnosing CD than IPSS. Current literature suggests the sensitivity of JVS ranges from 68.7% to 81.3% compared with 93.8% to 98% for IPSS (81, 83). There are several factors that improve JVS sensitivity including: (1) CRH stimulation; (2) positioning catheters against the medial walls of the jugular veins close to the IPS origins; and (3) performing a Valsalva maneuver during the procedure (to facilitate mixing of blood) (82). Erickson et al. enhanced JVS by adjusting the reference ratios for interpreting the results (81). They maximized the sensitivity (as described previously) using a pre-CRH IPS:P ACTH ratio of 1.59 and a post-CRH IPS:P ACTH ratio of 2.47. They observed, however, that during simultaneous JVS and IPSS, the former missed the diagnosis of CD in about 30% of cases (83). Arguably, JVS is less invasive and may be performed by less experienced radiologists. That said, if JVS must be substituted for IPSS, negative results (no central/peripheral gradient) should prompt referral to a tertiary center for IPSS confirmation. In most institutions, the only indication for JVS is unsuccessful IPS cannulation.

The utility of IPSS in pregnancy

The physiologic changes of pregnancy make testing for CS more difficult. The serum cortisol, plasma ACTH, corticosteroid-binding globulin and UFC are increased during the second and third trimesters of normal pregnancy, and response to dexamethasone is blunted (84, 85). There is a paucity of IPSS data during pregnancy because of concerns about fetal radiation exposure (86). Lindsay et al. reported IPSS results from 2 pregnant patents in whom the procedure was safely performed using special precautions including lead shielding and direct catheterization of the jugular veins. IPSS localized the tumor in both patients. The authors cautioned against the potential for false positives (elevated IPS:P ACTH ratio absent CD) during pregnancy because CS resulting from an adrenal lesion can be associated with nonsuppressed ACTH levels (85).

Other approaches to explore the etiology of ACTH-dependent CS

There are alternative strategies to evaluate the source of ACTH-dependent CS that may reduce the need for IPSS. Some studies have used dynamic testing (including the CRH stimulation test, desmopressin stimulation test, and high-dose dexamethasone suppression test) plus advanced imaging to distinguish between CD and EAS with variable results in the literature (87-91). A recent paper by Frete et al. assessed a diagnostic algorithm of CRH and desmopressin stimulation tests and pituitary MRI. The patients with inconclusive results underwent thin-slice whole-body CT scan. The combination of dynamic tests was 73% sensitive for CD. Using a combination of both dynamic tests and MRI scans, followed by CT scan if the diagnosis of CD was disputed, a PPV of 100% for CD was achieved. The authors concluded that IPSS could have been avoided in half of their patients (92).

A recent report in the Journal of Neurosurgery discussed a case of CD in which a 7-tesla (T) MRI was able to localize an otherwise invisible tumor. The authors suggest that 7-T imaging may preempt IPSS in standard and dynamic contrast 1.5-T and 3-T MRI-negative CD (93). This mirrors a previous study that was able to detect 3 pituitary adenomas on 7-T MRI unseen on 1.5-T imaging (94).

A study by Page-Wilson et al. evaluated the utility of neuroendocrine markers proopiomelanocortin and agouti-related protein levels in the diagnostic workup of ACTH-dependent CS. They demonstrated that values of proopiomelanocortin > 36 fmol/mL and agouti-related protein > 280 pg/mL yielded a sensitivity and specificity of 82% and a PPV of 100% for EAS (95).

Walia et al. looked at the use of Gallium-68 (⁶⁸Ga) tagged CRH during positron emission tomography (PET)-CT to evaluate the etiology of ACTH-dependent CS in 27 patients. The ⁶⁸Ga CRH PET-CT correctly delineated corticotropinoma in all 24 cases of CD, including 10 tumors of size < 6 mm. The location of the corticotropinoma on ⁶⁸Ga CRH PET fusion images with MRI scans were concordant with operative findings and confirmed on histopathology. In contrast, there was no pituitary tracer uptake in 2 patients with EAS. There was diffuse pituitary tracer uptake in another patient who was believed to have ectopic CRH production.

Conclusions

IPSS is the most sensitive and specific biochemical test to distinguish pituitary from ectopic ACTH-dependent CS. IPSS cannot be used to confirm the diagnosis of ACTH-dependent CS. There are certain criteria that need be met to ensure the accuracy, precision, and reliability of IPSS, which include the presence of hypercortisolism immediately before as well as at the time of the procedure. IPSS should be performed by an experienced interventional or neuroradiologist to avoid complications and nondiagnostic results. Samples must be collected and processed in a precise manner. Prolactin is an excellent marker to confirm proper IPS venous efflux if the IPS:P ACTH gradient suggests ectopic disease. The value of the intersinus ACTH gradient to predict tumor lateralization may be improved using the prolactin-adjusted ACTH ratio, but further study is needed. We hope that this case-based systematic review will help clinicians use a stepwise approach (Table 7) to interpreting IPSS results.

Table 7.

Important steps in performing and interpreting IPSS results

1. Confirmation of endogenous hypercortisolism before the procedure

2. Demonstrate hypercortisolemia before the procedure in patients with cyclical Cushing syndromea

3. Proper processing of blood samples (especially plasma ACTH)

4. Confirmation of adequate IPS venous flux using IPS:P prolactin ratios in patients who lack a significant IPS:P ACTH ratio

• The prolactin-adjusted IPS:P ACTH ratio can improve differentiation between Cushing disease and ectopic ACTH syndrome when there is a lack of proper IPS venous efflux based on IPS:P prolactin ratio

• An absolute IPS ACTH level < 200 and < 400 pg/mL pre- and post-CRH stimulation and a < 35% increase in ACTH-to-CRH in the periphery may suggest failed IPS cannulation

5. A lack of significant IPS:P ACTH gradient in unilateral successful IPS catheterization does not rule out a corticotroph adenoma to the contralateral gland

6. The value of the intersinus ACTH gradient to predict tumor lateralization may be improved by using a prolactin-adjusted ACTH ratio >1.4

Abbreviations: CRH, corticotropin-releasing hormone; IPS, inferior petrosal sinus; IPS:P, inferior petrosal sinus to peripheral ACTH gradient.

^a Some experts feel the duration of hypercortisolism before the IPSS may be more important than the presence of hypercortisolemia immediately before the procedure.

Glossary

Abbreviations

CD

Cushing disease

CRH

corticotropin-releasing hormone

CS

Cushing syndrome

CT

computed tomography

DDAVP

desmopressin

DST

dexamethasone suppression test

EAS

ectopic ACTH syndrome

IPS:P

inferior petrosal sinus sampling-to-peripheral gradient

IPSS

inferior petrosal sinus sampling

JVS

jugular vein sampling

MR

magnetic resonance

MRI

magnetic resonance imaging

NIH

National Institutes of Health

PET

positron emission tomography

PPV

positive predictive value

T

tesla

UFC

urinary free cortisol

68Ga

gallium 68

Additional Information

Disclosures: The authors have no conflicts to declare.

Data Availability

All data generated or analyzed during this study are included in this published article or in the data repositories listed in References. Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.