| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Endocrine Care |
Department of Endocrinology, St. Bartholomews Hospital, London EC1A 7BE, United Kingdom
Address all correspondence and requests for reprints to: Prof. Ashley Grossman, Department of Endocrinology, St. Bartholomews Hospital, London EC1A 7BE, United Kingdom. E-mail: . A.B.Grossman{at}qmul.ac.uk
Abstract
The CRH test is in widespread use for the differential diagnosis of ACTH-dependent Cushings syndrome (CS). Despite the greater availability worldwide of human-sequence CRH (hCRH), there are no large series reporting the response criteria that best discriminate between Cushings disease (CD) and the ectopic ACTH syndrome (EC) when using hCRH, rather than ovine-sequence CRH. We have, therefore, analyzed retrospectively the serum cortisol and plasma ACTH responses to hCRH in patients with ACTH-dependent CS, to develop response criteria that best discriminate between CD and EC.
One hundred fifteen consecutive patients with proven ACTH-dependent CS were studied: 101 with CD (78 females; mean age, 40 yr; range, 1073) and 14 with EC (7 females; mean age, 46 yr; range, 3269). The response to hCRH was also studied in 30 normal volunteers (NVs; mean age, 29 yr; range, 2044) with no clinical evidence of CS, and the results were compared. Following basal sampling at -15 and 0 min, hCRH (100 µg iv) was administered via an indwelling forearm cannula at 0900 h and serum cortisol and ACTH were measured at 15-min intervals for 2 h. The mean basal, peak, incremental, and percentage change in the serum cortisol and ACTH at all time points, and combination of time points, were calculated and analyzed to establish the best criteria to discriminate between CD and EC, and also between CD and NVs.
The mean serum cortisol concentration in samples obtained at 15 and 30 min after CRH increased by at least 14% above the mean basal in 85 of 100 patients with CD, but in none with EC, giving a sensitivity of 85% at a specificity set at 100%. In contrast, the best plasma ACTH response of a rise of 105%, calculated from the maximal rise, gave only 70% sensitivity at 100% specificity. In the NVs, the maximum cortisol at the mean 15+30 min time point was 615 nmol/liter. Using the 15 and 30 min time points as the reference point, 71 of 100 patients with CD had a rise of serum cortisol greater than 14% and also showed an absolute cortisol level more than 615 nmol/liter.
Serum cortisol responses to hCRH can be used to suggest the diagnosis of CD in the majority of patients with this condition, but it should only be used in conjunction with other biochemical and imaging modalities in establishing this important diagnosis. The measurement of plasma ACTH was less helpful in making this distinction, although it may have additional value in excluding ACTH-independent causes of CS. Although we believe that bilateral inferior petrosal sinus sampling remains the single most useful test in discriminating CD from EC in patients with ACTH-dependent CS, hCRH offers rapid diagnostic information and is a useful adjunctive test in establishing the presence of a possible ectopic source.
THE CRH TEST HAS been widely used as a noninvasive tool in the differential diagnosis of ACTH-dependent Cushings syndrome (CS) (1). Following administration of CRH the majority of patients with pituitary-dependent CS, Cushings disease (CD), show a rise in plasma ACTH and cortisol, usually to a peak level greater than in normal subjects, whereas those with nonpituitary sources of ACTH characteristically do not. The majority of series report the use of ovine-sequence CRH (oCRH) (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12): in the largest single series, this test had 93% sensitivity and 100% specificity in the discrimination of CD from ectopic ACTH secretion (EC) using appropriate criteria (13). Human-sequence CRH (hCRH) has qualitatively similar properties to oCRH, although in comparison with oCRH the human-sequence peptide elicits a smaller rise in plasma ACTH and cortisol in normal and obese individuals, and in those with CD (14). Previous small series have compared oCRH and the human-sequence peptide (hCRH) in the differential diagnosis of ACTH-dependent CS and concluded that oCRH is superior in this respect (15). Thus, reported response criteria from series using oCRH, or meta-analyses of published series where the results of both hCRH and oCRH are combined, are unlikely to be appropriate for hCRH testing. Current recommendations made by such analyses include a peak increment above basal of serum cortisol of more than 20%, or plasma ACTH of more than 50%, as being consistent with CD (16).
To date, there are no large series reporting the responses to hCRH in patients with ACTH-dependent CS, whereas hCRH is currently the predominant form of CRH in international use. Therefore, we have analyzed the responses seen during hCRH testing in a large series of patients with ACTH-dependent CS to develop criteria that best discriminate CD from EC. In particular, because transsphenoidal pituitary surgery is widely accepted as the primary treatment of CD, we have analyzed our data by setting the specificity for the diagnosis of CD at 100% such that inappropriate pituitary surgery is avoided in patients with EC. In addition, we have compared the responses seen in CD with those in a group of normal volunteers (NVs) to determine whether the hCRH test as a single investigation is also useful as a combined discriminator of CD from both normality and from EC.
Subjects and Methods
The results from every patient with ACTH-dependent CS presenting between 1987 (when hCRH became available for human testing) and 2000 were analyzed: 101 patients with CD (78 femalesmean age, 40 yr; median, 39; range, 1073) and 14 patients with EC (7 femalesmean age, 46 yr; median, 44; range, 3269). CD was confirmed in all cases by pathological confirmation of a pituitary adenoma immunostaining for ACTH, or clinical and biochemical cure (17) after selective microadenomectomy. In all patients with EC the diagnosis was confirmed by removal of a tumor showing positive ACTH immunoreactivity: six patients with bronchial carcinoid tumors, three patients with small cell lung cancer, two patients with pancreatic carcinoid tumors, two patients with medullary cell carcinoma of the thyroid, and one patient with a disseminated gut carcinoid. At the time of testing with hCRH, no patient was receiving any drug known to affect the hypothalamo-pituitary-adrenal axis. Twenty-three of these patients have previously been reported (18). In addition, 30 healthy NVs (all malemean age, 29 yr; median, 26; range, 2044) with no clinical evidence of CS were also studied. There is no gender variation in responsiveness to CRH in normal subjects (19).
After an overnight fast, an indwelling forearm cannula was inserted at 0830 h, and the patient remained supine for the remainder of the test. At 0900 h (0 min) 100 µg iv hCRH (Ferring Pharmaceuticals Ltd., Malmo, Sweden) was administered as a bolus dose over 30 sec. Blood was taken for estimation of serum cortisol and plasma ACTH at -15, 0, 15, 30, 45, 60, 90, and 120 min. Complete data sets of values at all time points were available for serum cortisol in 100 of 101 patients with CD, for plasma ACTH in 93 of 100 patients with CD, and for both serum cortisol and ACTH in all patients with EC. For the NVs, only serum cortisol was assessed. Patients with incomplete data sets were excluded from the analysis.
Calculation of response criteria
The use of the peak and peak incremental rise above a mean basal value at -15 and 0 min was assessed for serum cortisol and ACTH. In addition, the percentage rise above the mean basal level at each time point, and above each pair of time points, was calculated. Criteria were set to assess which values were best able to discriminate CD from EC, setting the specificity such that no patient with EC was incorrectly classified as having CD. A similar analysis was performed on the NV data.
Assays
After immediate centrifugation at 4 C of blood samples taken in cooled lithium-heparin tubes, plasma was immediately frozen and stored at -20 C until assay. Serum cortisol was measured by an in-house unextracted nonchromatographic RIA: the coefficient of variation at 100 nmol/liter and 1000 nmol/liter is 6% (20). Plasma ACTH was measured by our validated routine in-house Vycor (Socite-A.T.A., Geneva, Switzerland) glass-extracted RIA (21); inter- and intra-assay coefficients of variation are less than 8% for both.
Statistics
The group data are expressed as mean ± SEM. Repeated measures ANOVA was used to evaluate the time course of the hormonal response. Basal and peak ACTH and cortisol values between groups were compared using the Mann-Whitney U test. The level of significance was set at 5%.
Results
As a group, patients with CD showed a significant rise in both serum cortisol and plasma ACTH during the time course of the test (P < 0.001 for both), whereas those with EC did not (P > 0.1) (Fig. 1
). NVs showed a significant rise in serum cortisol (P < 0.001; data not shown).
|
Although the mean basal values for serum cortisol were significantly different between patients with CD and those with EC (P < 0.001), the range of basal values showed considerable overlap (Fig. 2
). Similarly, although the peak values were significantly different between the CD and EC groups, the peak values being higher in patients with EC (P < 0.05), there was considerable overlap, with only one patient with EC having a value greater than the highest value in the CD group (Fig. 2
). Calculation of the incremental change did not improve discrimination between groups (Fig. 2
). Thus, neither basal, peak, nor absolute incremental values are useful in discriminating CD from EC.
|
However, the best discrimination between the groups of CD and EC was achieved by calculating the percentage rise from mean basal serum cortisol level to the mean of values at 15 and 30 min, with a rise of 14% or more being seen in 85 of 100 patients with CD (sensitivity, 85%), but in no patient with EC (specificity, 100%) (Fig. 3
).
|
|
Plasma ACTH responses
The mean basal values for plasma ACTH were significantly different between patients with CD and those with EC (P < 0.001); however, there was no significant difference between the peak ACTH values between the two groups (P > 0.2). As with the serum cortisol responses, neither basal, peak, nor absolute incremental ACTH values were useful in discriminating CD from EC.
The best criterion was a maximal rise of plasma ACTH above basal of more than 105%, which was seen in 65 of 93 patients with CD, with levels below this in all EC patients, giving a 70% sensitivity at a specificity of 100% (Fig. 5
). No other time points alone or in combination gave superior discrimination. Analysis of the responses on a gender basis revealed that 57 of 72 women with CD showed a maximal rise in plasma ACTH of more than 87%, with all women with EC exhibiting a rise lower than this, giving better discrimination with a sensitivity of 79% at a specificity of 100%. Analysis of male responses showed the ACTH response to be a poorer discriminator than in women.
|
Adverse effects
After administration of hCRH, minor adverse effects were experienced in occasional individuals; these were limited to a short-lived flushing sensation and/or a brief metallic taste in the mouth occurring soon after injection of hCRH. No other problems were encountered.
Discussion
Discriminating between pituitary and ectopic sources of ACTH in ACTH-dependent CS remains, on occasion, a considerable clinical challenge (1). In women with ACTH-dependent CS there is a 10-fold higher prevalence of CD compared with EC, whereas in men the ratio is nearer to 3:1. Overall, 8590% of all cases of ACTH-dependent CS will be of pituitary origin. Here, we report that the best response criterion with hCRH testing is a rise above the mean basal in the mean serum cortisol at 15 and 30 min of 14% or more, with a sensitivity of 85% and specificity of 100% for the diagnosis of CD. We chose to maintain the specificity at 100% so that the test could be used as an aid to avoid inappropriate pituitary surgery. If a more cautious cutoff of a greater than 20% rise in serum cortisol at these time points is taken, the sensitivity falls slightly to 82%.
Examining the variability of the serum cortisol responses to hCRH in EC, the mean rise above basal was +2.7%. The upper 95% confidence limit for the responses to hCRH in EC is +12%. Thus, if a patient with ACTH-dependent CS shows a response to hCRH according to the criterion cited here (>14%), and assuming an a priori ratio of CD/EC of around 10:1, then there is a greater than 99% probability that the cause is CD rather than EC.
The largest single series of CRH testing originates from the NIH, reporting the results of the oCRH test in 100 patients with CD and in 16 with EC (13), similar to the numbers here. Using the best ACTH response criteria, their data indicated that the oCRH test had a sensitivity of 93% and a specificity of 100%. Their best cortisol response criterion was less discriminatory, indicating a sensitivity of 91% and a specificity of 88%. According to the data presented here, in the first instance there appeared to be no advantage in measuring ACTH in addition to cortisol, although ACTH would still be useful in differentiating ACTH-independent from ACTH-dependent causes. However, it does seem that hCRH may be slightly inferior to oCRH when used as a tool to differentiate between CD and EC. There are possible explanations for this. First, the response criteria reported by Nieman et al. (13) relied on basal blood samples drawn at -5- and -1-min time points that were not carried out in this study. They suggest that because plasma ACTH values may fluctuate rapidly in patients with ACTH-dependent disease, especially when of nonpituitary origin, and because plasma ACTH has a short half-life of about 10 min, samples drawn this close together may more accurately reflect basal values. It is possible that this might improve test accuracy where percentage responses are considered. Second, although hCRH administration causes a qualitatively similar plasma ACTH and serum cortisol response compared with oCRH, the duration of response and absolute increment is less, and this may reflect a more rapid clearance of the administered hCRH by endogenous CRH-binding protein (22). Therefore, the lower hCRH test sensitivity may be a reflection of this, and it is probably also the explanation of why the earlier time points for serum cortisol responses of 15 and 30 min, rather than 30 and 45 min as reported for oCRH, gave better discrimination for the human-sequence peptide.
The times at which the responses are calculated are important. Previous recommendations for responses that are consistent with CD include an incremental peak response of 20% for serum cortisol and more than 50% for plasma ACTH. If applied to our data, the sensitivity and specificity of the cortisol responses fell to 86% and 73%, and ACTH to 81% and 80%, respectively. Although we suggest that a 20% rise in serum cortisol as a criterion is strongly indicative of CD, this is calculated from the mean at 15 and 30 min. For the reasons discussed above, it is not appropriate to apply the same response criteria to the two different peptides. However, our data do suggest that continuation of the test beyond the 60-min sampling time is unnecessary.
Attempts have been made to improve on the sensitivity of the hCRH test. Because ACTH synthesis and secretion is also under the influence of vasopressin (VP) (23), combinations of lysine-VP and CRH have been used to improve the stimulation of ACTH and, hence, cortisol release (24, 25, 26). Even at low doses, however, infusion of lysine-VP may cause abdominal pain and nausea. Desmopressin has few side effects and, although inferior when used by itself for the differential diagnosis of ACTH-dependent CS (1, 27), when used in combination with hCRH the discrimination between EC and CD appeared to be complete (18). The numbers of patients studied were, however, small, and more data are required. Indeed, as noted above, it is unlikely that any dynamic test will offer 100% sensitivity and specificity for the differential diagnosis between CD and EC.
We have also compared the cortisol response in a group of NVs, to determine not only the best discrimination between CD and EC but also the validity of the test in determining between the cortisol response seen in the majority of patients with CD and the NVs. We have previously shown that responses to hCRH are comparable in NVs and patients with simple obesity (14). In clinical practice, patients falling into the latter group may frequently need to be excluded as suffering from CS. As may have been expected, absolute cortisol values rather than a percentage rise gives the best discrimination. Overall, discrimination based on the mean of values at 15 and 30 min being greater than 615 mmol/liter gives 80% sensitivity and 100% specificity. However, others have maintained that patients with simple obesity, particularly if this has an android or centripetal distribution, may have an enhanced response to CRH (28). Similarly, patients with severe depressive illness may also have marked cortisol responses to CRH, although usually still within the normal range (3). In our own recent study, we found the cortisol responses to oCRH to be superimposable with our normal controls (29), and this is likely also to apply to hCRH. Such patients may be those who are pseudo-Cushingoid and most need to be differentiated from CS. Our data suggest that hCRH may be of value in differentiating such pseudo-Cushingoid states from true CD. Using the hCRH test with measurement of serum cortisol, it may be noted that some 71% of our patients with CD showed response criteria excluding EC and with a maximal level at 15 + 30 min above that seen in our NVs.
By themselves, none of the noninvasive or dynamic tests used for the differential diagnosis of ACTH-dependent CS are completely reliable, and several tests may need to be used. If the hCRH test shows a rise above 20% in serum cortisol at 15 and 30 min, EC is very unlikely. However, the sensitivity of the test, in either its oCRH or hCRH form, for CD is still significantly less than 100%. What then are the recommendations for the use of the test in clinical practice? Bilateral inferior petrosal sinus sampling (BIPSS) remains the gold standard for establishing whether there is a central to peripheral gradient of plasma ACTH consistent with CD, with a sensitivity of 95100% with a specificity of virtually 100% in major centers (1). This is, however, an invasive test requiring considerable expertise in specialist centers, and some clinicians will only option for it in cases of doubt. Forty percent of patients with CD will have a normal pituitary magnetic resonance image, whereas conversely there is a 10% prevalence of pituitary incidentalomas in the age range in which CD typically presents (1). It is, therefore, clear that magnetic resonance imaging of the pituitary cannot be relied on and weight should be given to the biochemical evaluation of a given patient. With a clear response to hCRH and an obvious pituitary lesion with stalk deviation, BIPSS may not be needed. An additional advantage of the hCRH test is that if plasma ACTH is also measured a persistently undetectable ACTH will confirm ACTH independence and the primacy of adrenal pathology (1). However, given the problems in interpreting any dynamic test in isolation, we currently consider that BIPSS or cavernous sinus sampling will continue to be necessary in the great majority of patients.
Acknowledgments
We are grateful to all the staff on Francis Fraser and Garrod wards for performing the tests.
Footnotes
Present address for J.N.-P.: Division of Clinical Sciences, University of Sheffield, Northern General Hospital, Herries Road, Sheffield S5 7AU, United Kingdom.
Abbreviations: BIPSS, Bilateral inferior petrosal sinus sampling; CD, Cushings disease; CS, Cushings syndrome; EC, ectopic ACTH syndrome; hCRH, human-sequence CRH; NV, normal volunteer; oCRH, ovine-sequence CRH; VP, vasopressin.
Received May 25, 2001.
Accepted December 18, 2001.
References
This article has been cited by other articles:
![]() |
D. L. Batista, J. Riar, M. Keil, and C. A. Stratakis Diagnostic Tests for Children Who Are Referred for the Investigation of Cushing Syndrome Pediatrics, September 1, 2007; 120(3): e575 - e586. [Abstract] [Full Text] [PDF] |
||||
![]() |
R M Testa, N Albiger, G Occhi, F Sanguin, M Scanarini, S Berlucchi, M P Gardiman, C Carollo, F Mantero, and C Scaroni The usefulness of combined biochemical tests in the diagnosis of Cushing's disease with negative pituitary magnetic resonance imaging Eur. J. Endocrinol., February 1, 2007; 156(2): 241 - 248. [Abstract] [Full Text] [PDF] |
||||
![]() |
L. R. Salgado, M. C. B V. Fragoso, M. Knoepfelmacher, M. C. Machado, S. Domenice, M. A. A. Pereira, and B. B. de Mendonca Ectopic ACTH syndrome: our experience with 25 cases. Eur. J. Endocrinol., November 1, 2006; 155(5): 725 - 733. [Abstract] [Full Text] [PDF] |
||||
![]() |
R.T. Netea-Maier, W.-A. Nieuwlaat, C.G.J. Sweep, P. Wesseling, L. Massuger, and A.R.M.M. Hermus Virilization due to ovarian androgen hypersecretion in a patient with ectopic adrenocorticotrophic hormone secretion caused by a carcinoid tumour: Case Report Hum. Reprod., October 1, 2006; 21(10): 2601 - 2605. [Abstract] [Full Text] [PDF] |
||||
![]() |
R T Netea-Maier, E J van Lindert, M den Heijer, A van der Eerden, G F F M Pieters, C G J Sweep, J A Grotenhuis, and A R M M Hermus Transsphenoidal pituitary surgery via the endoscopic technique: results in 35 consecutive patients with Cushing's disease. Eur. J. Endocrinol., May 1, 2006; 154(5): 675 - 684. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. M. Isidori, G. A. Kaltsas, C. Pozza, V. Frajese, J. Newell-Price, R. H. Reznek, P. J. Jenkins, J. P. Monson, A. B. Grossman, and G. M. Besser The Ectopic Adrenocorticotropin Syndrome: Clinical Features, Diagnosis, Management, and Long-Term Follow-Up J. Clin. Endocrinol. Metab., February 1, 2006; 91(2): 371 - 377. [Abstract] [Full Text] [PDF] |
||||
![]() |
I. Ilias, D. J. Torpy, K. Pacak, N. Mullen, R. A. Wesley, and L. K. Nieman Cushing's Syndrome Due to Ectopic Corticotropin Secretion: Twenty Years' Experience at the National Institutes of Health J. Clin. Endocrinol. Metab., August 1, 2005; 90(8): 4955 - 4962. [Abstract] [Full Text] [PDF] |
||||
![]() |
Y. S. Woo, A. M. Isidori, W. Z. Wat, G. A. Kaltsas, F. Afshar, I. Sabin, P. J. Jenkins, J. P. Monson, G. M. Besser, and A. B. Grossman Clinical and Biochemical Characteristics of Adrenocorticotropin-Secreting Macroadenomas J. Clin. Endocrinol. Metab., August 1, 2005; 90(8): 4963 - 4969. [Abstract] [Full Text] [PDF] |
||||
![]() |
H. L Storr, F. Afshar, M. Matson, I. Sabin, K. M Davies, J. Evanson, P N. Plowman, G M. Besser, J. P Monson, A. B Grossman, et al. Factors influencing cure by transsphenoidal selective adenomectomy in paediatric Cushing's disease Eur. J. Endocrinol., June 1, 2005; 152(6): 825 - 833. [Abstract] [Full Text] [PDF] |
||||
![]() |
H. L. Storr, A. M. Isidori, J. P. Monson, G. M. Besser, A. B. Grossman, and M. O. Savage Prepubertal Cushing's Disease Is More Common in Males, But There Is No Increase in Severity at Diagnosis J. Clin. Endocrinol. Metab., August 1, 2004; 89(8): 3818 - 3820. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. Arnaldi, A. Angeli, A. B. Atkinson, X. Bertagna, F. Cavagnini, G. P. Chrousos, G. A. Fava, J. W. Findling, R. C. Gaillard, A. B. Grossman, et al. Diagnosis and Complications of Cushing's Syndrome: A Consensus Statement J. Clin. Endocrinol. Metab., December 1, 2003; 88(12): 5593 - 5602. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. M. Isidori, G. A. Kaltsas, S. Mohammed, D. G. Morris, P. Jenkins, S. L. Chew, J. P. Monson, G. M. Besser, and A. B. Grossman Discriminatory Value of the Low-Dose Dexamethasone Suppression Test in Establishing the Diagnosis and Differential Diagnosis of Cushing's Syndrome J. Clin. Endocrinol. Metab., November 1, 2003; 88(11): 5299 - 5306. [Abstract] [Full Text] [PDF] |
||||
![]() |
H. L. Storr, P. N. Plowman, P. V. Carroll, I. Francois, G. E. Krassas, F. Afshar, G. M. Besser, A. B. Grossman, and M. O. Savage Clinical and Endocrine Responses to Pituitary Radiotherapy in Pediatric Cushing's Disease: An Effective Second-Line Treatment J. Clin. Endocrinol. Metab., January 1, 2003; 88(1): 34 - 37. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. A. Kelly, G. Samandouras, A. B. Grossman, F. Afshar, G. M. Besser, and P. J. Jenkins Neurosurgical Treatment of Nelson's Syndrome J. Clin. Endocrinol. Metab., December 1, 2002; 87(12): 5465 - 5469. [Abstract] [Full Text] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |