Article Text

Download PDFPDF

Original research
Bilateral inferior petrosal sinus sampling: experience in 327 patients
  1. Amy Deipolyi1,
  2. Alexander Bailin2,
  3. Joshua A Hirsch3,
  4. T Gregory Walker2,
  5. Rahmi Oklu4,5
  1. 1Division of Vascular & Interventional Radiology, Department of Radiology, New York University Medical Center, New York, New York, USA
  2. 2Massachusetts General Hospital, Harvard Medical School, Vascular & Interventional Radiology, Boston, Massachusetts, USA
  3. 3Division of Interventional Neuroradiology, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, USA
  4. 4Mayo Clinic, Vascular & Interventional Radiology, Scottsdale, Arizona, USA
  5. 5Department of Medicine, Biomaterials Innovation Research Center, Brigham and Women's Hospital, Harvard Medical School, Cambridge, Massachusetts, USA
  1. Correspondence to Rahmi Oklu, Mayo Clinic, Division of Interventional Radiology, 13400 E Shea Blvd, Scottsdale, AZ 85259, USA; oklu.rahmi{at}


Objective To describe findings and outcomes of 331 bilateral inferior petrosal sinus sampling (BIPSS) procedures performed in 327 patients evaluated for Cushing disease (CD).

Materials and methods The radiology department's electronic database was searched to identify all BIPSS procedures (1990–2013). Electronic medical records were used to identify demographics, laboratory, procedural, surgical and pathologic findings.

Results A total of 331 BIPSS procedures were performed in 327 patients (254 F, 73 M), mean age 41 (range 7–81) years. The overall technical success rate was 88% for bilateral cannulation, though nearly two-thirds of the technical failures had unilateral sampling that diagnosed CD. Of the 331 BIPSS procedures, 40 were performed without, and 291 with stimulation by Acthrel or desmopressin. Sensitivity was 89–94% for unstimulated BIPSS, 96% for stimulated BIPSS, and 77% for MRI. BIPSS lateralization was accurate in about half of patients, compared with 75% accuracy for MRI. Mean inferior petrosal sinus (IPS):peripheral adrenocorticotropic hormone ratio was 17.3 (SE 1.8) at baseline, and 99.2 (SE 14.8) at 3 min, with decreasing values over time. All patients with follow-up after surgical resection for centralizing BIPSS were reported to be cured, with cortisol levels significantly decreased from 19 to 4 μg/dL (p<0.0001). Complications from BIPSS were rare, including groin hematoma (2.5%), but no thromboembolic complications were seen.

Conclusions BIPSS remains the ‘gold standard’ for diagnosing CD. Stimulation with Acthrel or desmopressin is key to increasing specificity. When only one IPS can be successfully cannulated, results may still be diagnostic. BIPSS findings cannot be used to accurately lateralize lesions within the pituitary.

  • Angiography
  • Catheter
  • Technique
  • MRI
  • Tumor

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Bilateral inferior petrosal sinus sampling (BIPSS) is a minimally invasive test performed by interventional radiologists to evaluate challenging cases of Cushing syndrome (CS). A cause of CS is Cushing disease (CD), which is hypercortisolemia due to excess pituitary secretion of adrenocorticotropic hormone (ACTH), which stimulates adrenal cortical secretion of cortisol. CD is rare and diagnosis may be elusive. CD must be distinguished from ectopic sources of ACTH; non-invasive tests have relatively low sensitivity and specificity.1 ,2 Because surgery carries significant morbidity,3 performing a minimally invasive procedure that is safe to demonstrate central-peripheral gradient and justify a risky surgical procedure is often an appropriate diagnostic step before resection.

BIPSS can be accurate and is helpful, particularly in patients without definitive imaging findings. We evaluated the experience at a tertiary care center in 327 patients. Historically, the technique has undergone significant changes and improvements over the past few decades, and today it is safer, more reliable, and offers useful data that can be used to determine whether surgery is warranted, particularly in patients with equivocal laboratory assessment. Non-invasive imaging such as MRI, which can be helpful, may not provide a definitive diagnosis. There is a 10–20% incidence of identifying an incidental non-functioning lesion; additionally, the lesions may be too small to detect by MRI.1 ,4 We describe the accuracy and safety of BIPSS in a large series of patients to characterize the role of this procedure in evaluating patients with suspected CD.

Materials and methods

This institutional review board (IRB)-approved study, compliant with the Health Insurance Portability and Accountability Act (HIPAA), involved a search of the radiology department's electronic database to identify all BIPSS procedures between 1990 and 2013. Electronic medical records were used to identify demographics, clinical history, endocrine test results, surgical and pathology findings, and follow-up results for each patient.

BIPSS was performed as previously described,5 with routine intraprocedural heparinization. Both common femoral veins were accessed, with a six French sheath on the right and a five French sheath on the left. The inferior petrosal sinus (IPS) was selected bilaterally with five French catheters parked in the internal jugular veins, and a microcatheter positioned in the IPS and confirmed with hand injections. Before the introduction of Acthrel (synthetic corticotropin-releasing hormone; 1 µg/kg; Ferring Pharmaceuticals, Parsippany, New Jersey, USA), samples were collected from both IPSs and the periphery via the right sheath at various time points, placed on ice, and unstimulated specimens sent for assay of ACTH. During later procedures, samples were collected at baseline and multiple time points after administration of Acthrel or desmopressin acetate (DDAVP, synthetic vasopressin), stored on ice, and then assayed for ACTH. Samples were not collected from the internal jugular veins. The test was considered diagnostic of CD if baseline or unstimulated ACTH levels from the IPS were two times greater or more than from the periphery, or if stimulated specimens were three times greater or more than from the periphery. The study was furthermore considered lateralizing if the intersinus ratio was ≥1.4.2 ,6

Statistical tests included paired t tests to evaluate cortisol levels before BIPSS and after surgical resection. Repeated measures one-way analysis of variance with Tukey post hoc tests corrected for multiple comparisons were used to assess the changes in IPS:peripheral ACTH ratios from baseline over multiple time points during BIPSS.


Between 1990 and 2013, 331 BIPSS procedures were performed in 327 patients (254 F, 73 M), with an average age of 41 (range 7–81) years. The most commonly reported symptoms included weight gain (85%), moon facies (76%), skin changes (49%), sex hormone changes (41%) hypertension and/or diabetes (37%), muscle and bone weakness (34%), fatigue (25%), and mood and cognitive changes (24%) (table 1).

Table 1

Presenting symptoms in all patient who underwent bilateral inferior petrosal sinus sampling

Of the 331 BIPSS procedures, 40 were performed without, and 291 were performed with, stimulation by Acthrel or desmopressin. In 4 of the unstimulated and 36 of the stimulated BIPSS procedures, both IPSs could not be cannulated, giving an overall technical success rate of 88% for bilateral cannulation among 291 total cases. Most commonly, variant anatomy and/or inability to cannulate one or both IPSs were cited as the reason for failure. In one of the unstimulated BIPSS procedures, the left IPS was successfully cannulated and had ACTH levels greater than twice that of the periphery, and was therefore centralizing to the pituitary. Among the 36 technically unsuccessful stimulated BIPSS procedures, 19 had successful cannulation of one IPS with failure to cannulate the contralateral IPS. Of these, 12 (63%) had values from the successfully cannulated side exceeding the threshold of three times or greater ACTH value compared with the periphery, consistent with a centralized pituitary source.

Among the remaining 291 technically successful cases of BIPSS, 43 had incomplete medical records, were lost to follow-up, or did not have pathologic confirmation of the disease process. Among these 43 studies, 26 (60%) were centralizing and 17 suggested an ectopic ACTH source. Owing to lack of follow-up, these 43 studies were excluded from further analysis, leaving 26 unstimulated and 222 stimulated cases. Of the 248 cases, 17 (7%) were non-centralizing and suggested ectopic sources of ACTH; 6 of these went on to have pituitary resection and were confirmed not to have pituitary ACTH sources; the other 11 did not undergo surgery. Unstimulated BIPSS resulted in sensitivity of 94%. Baseline BIPSS results in the stimulated protocol yielded 89% sensitivity. Stimulated BIPSS resulted in sensitivity of 96%. Specificity could not be accurately calculated because many of the non-centralizing BIPSS patients did not undergo subsequent resection to provide pathologic diagnosis. Of the stimulated cases, 182 were pathologically confirmed and centralized to the pituitary, consistent with CD. Of the 182 cases, 14 (8%) were not centralized at baseline, but became centralized after stimulation.

Among the 182 patients with pathologic confirmation of CD, 160 (88%) had lateralization suggested by an intersinus ratio of ≥1.4, and 22 (12%) had no lateralization suggested. Among the 22 non-lateralizing studies, 18 (82%) had lateralization seen on pathology, with only four centrally located adenomas identified. For lateralizing BIPSS, 83/160 (52%) were concordant in laterality with pathology; 42/160 (26%) lateralized to the contralateral side; 35/160 (22%) lateralized to one side of the pituitary but were found to be centrally located on pathology. Therefore, only 87 (4 plus 83) of 182 BIPSS studies (48%) were accurate in predicting lateralization on pathology.

Of the patients with surgical pathology, 119 also had an MRI before BIPSS. MRI carried a 77% sensitivity. Of the 119 with an MRI, 85 had findings reported on MRI and surgical pathology after resection. Of these, 18 were described as central within the pituitary on MRI. Eight of these 18 (44%) were central within the pituitary on pathology. The remaining 67 MRI reports indicated a side of abnormality; 56 of these (84%) were concordant with pathology. Thus, there was an overall accuracy in lateralization by MRI of 75%.

Data from the 182 patients with centralizing BIPSS with pathologic confirmation were used to evaluate changes in ACTH level over time during sampling. For patients with non-lateralizing results, levels from both IPSs were averaged. For lateralizing studies, the levels were used from the side suggested by BIPSS. The two baseline levels were averaged. Mean IPS:peripheral ACTH ratio was 17.3 (SE 1.8) at baseline, and 99.2 (SE 14.8) at 3 min, with decreasing values over time (figure 1). Repeated measures one-way analysis of variance shows that the changes in the ratio over time are significant (F=4.18, p<0.0001); Tukey's multiple comparison post hoc test demonstrated significant differences between baseline and all time points (p<0.05) except 15 min (p>0.05), between 3 min and all times other time points (p<0.01), and between 5 min and 15 min (p<0.001). The values collected at 10 min did not differ from those collected at 5 or 15 min (p>0.05).

Figure 1

Adrenocorticotropic hormone (ACTH) values during stimulated bilateral inferior petrosal sinus sampling for patients with pathology-proven Cushing disease. ACTH levels in the lateralized inferior petrosal sinus were divided by the ACTH levels in the peripheral specimen, with the highest ratios seen at 3 min.

Complications from BIPSS were rare, and included groin hematoma (2.5%), which resolved without further management. No thromboembolic complications were noted. Among patients with centralizing BIPSS, mean cortisol levels before BIPSS were 19 μg/dL and after resection were 4 μg/dL (t=13.3, p<0.0001). Among patients with non-centralizing BIPSS who underwent pituitary resection, mean cortisol was 23 μg/dL before BIPSS and 10 μg/dL after surgery (p=0.0007).


BIPSS is a safe procedure with a high technical success rate. In this large series documenting the 331 BIPSS procedures in 327 patients, we report sensitivity of 96%. Findings are similar to those of other studies of moderately sized cohorts of 100–200 patients, demonstrating sensitivity exceeding 90%.7–9 No thromboembolic complications, the most dreaded adverse events following BIPSS, were found in this study, probably owing to a protocol that included routine heparinization.2 ,4 Of patients with centralizing BIPSS with pathologic confirmation and clinical follow-up, all were noted to have a clinical cure, with significantly decreased cortisol levels.

With the incorporation of stimulation with corticotropin-releasing hormone or desmopressin, sensitivity increased, corroborating the findings of previous reports.10–13 Furthermore, 8% of the BIPSS procedures that were not centralizing at baseline became centralizing after stimulation, enforcing the importance of hormone stimulation. While Acthrel was most frequently used, desmopressin may be used safely and effectively in its place, which was necessary during the study period owing to a Acthrel shortage.12 The study therefore reaffirms the importance of including stimulation in routine BIPSS protocols.

Analysis of IPS:peripheral ACTH ratios before and at multiple time points after stimulation showed that the maximum increases in ratios were seen at 3 min. Findings are similar to a prior study demonstrating maximal response at 3–5 min.8 In our study, there were no differences in ratios between baseline and 15 min. Therefore, we would suggest that this last time point can probably be excluded during routine BIPSS.

For pathology-proven cases, we found that BIPSS was consistent with pathologic lateralization in about half of cases. Findings here are consistent with prior reports of the relatively low accuracy of BIPSS in lateralizing CD, and are probably due to the collateral venous drainage pathways of the petrosal sinus.2 ,14 Because of this inaccuracy, surgeons are obliged to explore the entire pituitary after a centralizing BIPSS is demonstrated. Though MRI was less sensitive than BIPSS in detecting pituitary pathology, when a lesion is appreciated on imaging, the lateralization is more accurate (75%). Therefore, MRI findings may be helpful in combination with centralizing BIPSS results to assist in surgical planning.

Most technical failures are due to an inability to cannulate the IPS, most commonly due to venous anatomic variants or occlusion of the internal jugular vein.15 The reason for failed cannulation was not evident from the medical records, but probably relates to the 25% rate of drainage via a plexus, which may preclude cannulation with a microcatheter.4 In one case, the right internal jugular vein was found to be occluded; to perform the procedure, direct puncture of the internal jugular vein just above the occluded segment was performed to allow for cannulation of the IPS. Because of the high technical success rate in the study, at our institution, we have not routinely assessed prolactin levels as a positive control for IPS cannulation.16 ,17

In half of the technical failures in this study, one IPS was successfully cannulated. Of these, nearly two-thirds demonstrated centralizing IPS:peripheral ACTH ratios. Given the lack of consistency in BIPSS lateralization and the extensive venous collateralization, such a centralizing study can be considered diagnostic of CD. The negative studies, however, cannot be considered true negatives as there might be a contralateral lesion with venous drainage that was not successfully sampled. One IPS cannot be cannulated from the contralateral IPS, as there is no direct venous connection between the IPSs, but rather a venous plexus. However, gentle hand injection from one IPS that is successfully cannulated will demonstrate the contralateral IPS with reflux, which may be used as a guide for contralateral cannulation in difficult cases. In challenging cases, we have recently started to perform rotational angiography via injection of 3–5 cc of 1:2 diluted contrast over 6 s; this technique may be more helpful in providing a visual proof of IPS catheterization, rather than awaiting postprocedure prolactin levels (figure 2).

Figure 2

Rotational cone-beam CT to assure inferior petrosal sinus (IPS) catheterization. (A) DSA image demonstrates gentle hand injection of the left IPS (black arrow indicates microcatheter tip), with filling of the cavernous sinus (white arrowhead) and subsequently contralateral filling of the right IPS (white arrow). (B) Corresponding reformatted coronal images from the rotational cone-beam CT demonstrate the left microcatheter tip (black arrow), filling of the cavernous sinus (white arrowhead) and reflux into the contralateral IPS (white arrow). (C) Magnified view of panel B.

The primary limitation of the study is its retrospective nature, which may introduce bias in regard to which patients undergo BIPSS and, ultimately, surgery. Because most patients with non-centralizing BIPSS do not undergo surgical resection, there is no pathology available to calculate the specificity accurately, as the number of true negatives is unknown. Also, using pathology as a gold standard to assess BIPSS performance may be misleading, as some cases may not be definitive. Of the 25 patients with centralizing BIPSS who underwent surgical resection and were not found to have an adenoma on pathology, 17 (68%) obtained clinical cures, highlighting the potential limitation of pathologic studies. Finally, because many patients were referred from outside institutions, several patients had incomplete records and were excluded; other patients’ prior imaging results were also not available.

In conclusion, BIPSS is a safe and accurate test used in the investigation of patients with CS. BIPSS should be performed with Acthrel or desmopressin stimulation and with routine intraprocedural heparinization. Although BIPSS lateralization is not accurate owing to collateral venous drainage, sampling from one IPS when contralateral cannulation is not possible may be helpful and diagnostic. BIPSS remains the gold standard for diagnosing a pituitary source of ACTH in patients with hypercortisolemia.



  • Contributors AD performed data collection and analysis, and wrote the manuscript. AB designed data collection tools, collected the data, and revised the manuscript. JAH and GTW wrote and revised the paper. RO conceived the study design, oversaw data collection, and wrote and revised the manuscript; he is the guarantor.

  • Competing interests None declared.

  • Ethics approval Institutional review board.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement Happy to share all aspects of the study.