Evaluation of the clonidine suppression test in the diagnosis of phaeochromocytoma

Abstract

The aim of this study is to review the experience of the clonidine suppression test in a regional endocrine centre and to compare the diagnostic sensitivity and specificity using various previous published criteria. The design used is retrospective study. The subjects include 56 patients in whom clonidine suppression tests had been performed from 1995 to 2000: 15 with phaeochromocytoma and 41 patients in whom the diagnosis was excluded using a combination of biochemical testing, abdominal computed tomography scanning and clinical follow-up. Plasma catecholamines were measured by high pressure liquid chromatography on basal samples and at hourly intervals for 3 h after the administration of clonidine 300 μg orally and the following diagnostic criteria were applied: plasma noradrenaline+adrenaline>2.96 nmol l–1 at 3 h post-clonidine or a baseline plasma adrenaline plus noradrenaline>11.82 nmol l–1; plasma noradrenaline>2.96 nmol l–1 at 3 h post-clonidine and plasma noradrenaline>2.96 nmol l–1 and <50% fall in noradrenaline at 3 h post-clonidine. The results obtained is that mean plasma noradrenaline plus adrenaline fell across the test in 40/41 patients in the non-phaeochromocytoma patients and was lowest at 3 h (basal 2.28±0.14 vs 1.36±0.11 nmol l–1, P<0.001). In the phaeochromocytoma group, clonidine had a variable effect on adrenaline plus noradrenaline levels with increases in 7/15. Using an abnormal result as a 3 h level of noradrenaline plus adrenaline>2.96 mmol l–1 gave a sensitivity of 93% and specificity of 95%. When a 3 h noradrenaline>2.96 mmol l–1 was used, sensitivity was 87% and specificity 95%. Using the former criteria, noradrenaline plus adrenaline>2.96 mmol l–1, 1/15 in the phaeochromocytoma group had a normal result after clonidine suppression testing. Two of 41 in the non-phaeochromocytoma group had a false-positive result. Under carefully controlled conditions, the clonidine suppression test is well tolerated, safe and accurate for use in the investigation of patients with suspected phaeochromocytoma.

Introduction

Phaeochromocytomas are rare catecholamine-producing tumours that arise from chromaffin cells. Most are situated within the adrenal medulla; however, between 9 and 23% develop from extra-adrenal chromograffin tissue and are referred to as paragangliomas.1 Diagnosis is important as, if untreated, they can lead to fatal cardiovascular complications.2 They should be considered in patients with sustained or paroxysmal hypertension especially if accompanied by headache, palpitations (with or without tachycardia), sweating, pallor, tremor, extreme anxiety and hyperglycaemia.3, 4 A history of phaeochromocytoma, evidence of neurofibromatosis type 1, multiple endocrine neoplasia type 2A or 2B, von Hippel Lindau disease, succinate dehydrogenase complex subunit B or succinate dehydrogenase complex subunit D or a family history of any of the above should also alert the clinician to search for a phaeochromocytoma.5

The key to biochemical diagnosis is demonstration of elevated levels of plasma and urinary catecholamines and/or their metabolites (metanephrines). Difficulties can arise if 24 h urinary collections are incomplete, resulting in falsely low total catecholamine or metanephrine measurements. It may be helpful, however, to measure the concentration of catecholamines or metabolites per gram of creatinine if complete collection of urine is a concern. As with all biochemical tests, false positives must be expected and, therefore, high levels of catecholamines and their metabolites do not necessarily reflect the presence of a tumour.6 Patients with essential hypertension may have suggestive symptoms and borderline increases in catecholamines or, alternatively, drugs can interfere with the laboratory assays thus creating diagnostic uncertainty.7 Such drugs, which will be discussed later, should be withheld for 1–2 weeks before testing.

The clonidine suppression test was introduced by Bravo et al.8 to address the problems outlined above. It is based on the principle that clonidine suppresses normal neurogenically mediated catecholamine release, but does not affect autonomous tumour production of catecholamines. It is, therefore, useful in distinguishing patients with essential hypertension, who have borderline increases in catecholamines and some suggestive symptoms, from those who actually have a phaeochromocytoma.7 We wish to clarify, however, that 10 of our 15 patients who had phaeochromocytoma did not have catecholamines four times the upper limit of normal and our experience was not large enough for us to comment on this as a definitive diagnostic test without further evaluation. Approximately 4% of incidentalomas are phaeochromocytomas and, therefore, the presence of an incidentaloma on imaging is another important reason for the clinician to search for pheochromocytoma. Several criteria have been proposed for interpretation of the test in order to achieve maximum sensitivity and specifity.8, 9, 10 These are set out in Methods.

We have reviewed our own experience of the clonidine suppression test. Our patient base is within a regional endocrinology and hypertension referral centre in a closely defined clinical population of 1.6 million. Very few patients are referred outside the region without being assessed here initially, and virtually all phaeochromocytomas diagnosed are treated within our centre. We wished to identify and refine optimal criteria for interpretation of the test by comparing the diagnostic sensitivity and specificity using the various criteria previously proposed in the literature.

Methods

We reviewed the charts of all patients who underwent clonidine suppression tests in the Regional Centre for Endocrinology and Diabetes, Royal Victoria Hospital, from 1995 to 2000. This comprised 56 patients with suspected phaeochromocytoma (male:female 41:15, mean age 52±3.2 s.e.m.) based on their symptomatology and/or previously abnormal laboratory findings, namely elevated urinary catecholamines. Of these patients, 15 were shown as having phaeochromocytoma, which was proven histologically in all cases. In this time frame, an additional six patients with phaeochromocytoma went straight to surgery without clonidine suppression testing. Three had recurrent tumours and presented with typical symptoms, markedly high urinary catecholamines and computed tomography abnormalities in keeping with phaeochromocytoma. One had an extra-adrenal (bladder) tumour, which was diagnosed histologically post-surgery. Two others had typical symptoms, markedly raised urinary catecholamines and large tumours typical of phaeochromocytoma on computed tomography scanning. Therefore, over 5 years, 21 patients were treated for the condition. The remaining 41 patients had the diagnosis excluded by a combination of biochemical and pharmacological tests, abdominal computed tomography and clinical follow-up of at least 1 year. Several patients have ongoing follow-up for their primary condition, either at diabetic or hypertension clinic.

As well as our clinical follow-up, we also liaised with the state pathologist for Northern Ireland who has consulted the regional database for deaths from phaeochromocytoma. There was one death between 1984 and 2007 from undiagnosed phaeochromocytoma. This was of a 51-year-old female who was not known to our centre.

The clonidine suppression tests were performed as per the detailed protocols described by Bravo et al. All antihypertensive medications were withdrawn 48 h before the test. Patients were admitted the evening before and fasted from 2200 hours. The studies were performed in a quiet, warm side room with the patient remaining supine throughout the study period. Following insertion of the intravenous cannula, the patient rested in the supine position for 30 min. Blood pressure and heart rate were measured three times at 5 min intervals. Blood samples were taken twice, at 5 min intervals, for measurement of plasma adrenaline and noradrenaline. At time zero, clonidine 0.3 mg was administered orally in water. Blood pressure and heart rate were recorded at 30 min intervals and blood was taken at hourly intervals for plasma catecholamine levels until 3 h after clonidine administration. The frequency of sampling has since been altered in our protocol as basal and 3 h post-clonidine levels only are required.

Each blood sample was placed in a sodium metabisulphide lithium heparin tube and transported immediately to the laboratory on ice, centrifuged and stored at −20 °C until analysis. Measurements of plasma catecholamine levels were by high pressure liquid chromatography11, 12 with electrochemical detection.

The following criteria from the literature were used. Tests were positive if

criterion 1: plasma noradrenaline >2.96 nmol l–1 at 3 h post-clonidine;8

criterion 2: plasma noradrenaline plus adrenaline >2.96 nmol l–1 at 3 h post-clonidine or a baseline plasma adrenaline plus plasma noradrenaline >11.82 nmol l–1;9

criterion 3: plasma noradrenaline >2.96 nmol l–1 at 3 h after clonidine and <50% fall in noradrenaline compared with basal level at 3 h post-clonidine.10

Results

Clonidine was well tolerated by all patients. Most of them slept throughout the study and all were subsequently able to resume their normal activities. Blood pressure and heart rate fell in all patients, but none had hypotensive symptoms. Response of blood pressure and heart rate was not useful in delineating patients with from those without phaeochromocytoma (Table 1).

Table 1 Cardiovascular and plasma catecholamine responses to oral clonidine

Using the first criterion for diagnosis in which the clonidine suppression test is positive when the plasma noradrenaline is >2.96 nmol l–1 at 3 h gave a sensitivity of 87% (13/15 patients) and a specificity of 95% (39/41) (Figure 1).

Figure 1
figure1

Plasma noradrenaline values in individual patients before and 3 h after a single oral dose of clonidine.

The sensitivity of the test was improved when criterion 2 was used and a positive test confirmed when the plasma noradrenaline plus adrenaline were >2.96 nmol l–1 at 3 h. Thus, the sensitivity increased to 93% (14/15 patients) and specificity was 95% (Figure 2). Furthermore, using the alternative for criterion 2, 6 of the 56 patients had baseline noradrenaline plus adrenaline levels >11.82 nmol l–1 and all were confirmed to have phaeochromocytoma. It could be argued that patients in whom diagnosis was based on this criterion do not actually require clonidine suppression testing. Patients are often referred to our Regional Centre for Endocrinology and Diabetes with abnormal urinary catecholamines. In these cases, referring physicians are usually looking for a definitive answer, either confirming or refuting the diagnosis of phaeochromocytoma. We have chosen to go straight to clonidine testing early in these cases and it has served us well. In more local patients in other centres, it would be possible to look at basal plasma catecholamines first before proceeding to clonidine suppression testing.

Figure 2
figure2

Plasma total noradrenaline + adrenaline in individual patients before and 3 h after a single oral dose of clonidine.

In the patients in whom phaeochromocytoma was not found, application of the criterion of plasma noradrenaline plus adrenaline falling to <2.96 nmol l–1, indicating exclusion of the diagnosis, occurred in 40 of these 41 patients during the test and was lowest at 3 h. The degree of suppression was closely related to basal levels of noradrenaline and adrenaline (Figure 2). Consequently, using criterion 3 (above), which is based on a 50% fall in noradrenaline from baseline, resulted in a significant false-positive rate in patients with already low baseline noradrenaline with a sensitivity of 87% and specificity of 32%.

Comparisons of the sensitivities and specificities of the three criteria are shown in Table 2.

Table 2 Sensitivity and specificity of clonidine suppression test results using varying diagnostic thresholds

Discussion

In the majority of patients with typical signs and symptoms of phaeochromocytoma, the demonstration of markedly increased levels of plasma or urinary catecholamines and their metabolites strongly suggest the presence of a phaeochromocytoma. Evidence shows that in the absence of conditions that can cause significant stress, plasma noradrenaline plus adrenaline of >11.82 nmol l–1 (2000 ng l–1) is diagnostic of a phaeochromocytoma.9, 13 Six of the 15 patients with phaeochromocytoma fulfilled this criterion, whereas the 41 patients without phaeochromocytoma did not.

When biochemical tests are equivocal, it is necessary to distinguish those patients with borderline increases in catecholamines from those with underlying phaeochromocytoma. Before additional tests are used, consideration should be given to possible causes of false-positive results and this was completed in our cohort. Certain clinical conditions may increase both plasma and urinary catecholamines and their metabolites to levels usually seen in phaeochromocytoma. These include acute clonidine withdrawal, acute alcohol withdrawal, vasodilator therapy with hydralazine or minoxidil, acute myocardial ischaemia or infarction, acute cerebral vascular accident, cocaine abuse, acidosis, hypoglycaemia, hypotension in shock and severe congestive heart failure.8 Diet can also contribute to false-positive result. For example, caffeic acid, a catechol found in coffee, is known to interfere with the assays of plasma catecholamines.14 This effect was eliminated in our patients as those undergoing testing were fasted overnight.

Biochemical testing for phaeochromocytoma should ideally be carried out after discontinuation of medications known to elevate levels of catecholamines and their metabolites or interfere directly with biochemical analyses. Antidepressants alter the alpha-2-imidazole autoreceptor sensitivity, eliminating the clonidine suppressive effects,7 whereas beta-blockers can inhibit the metabolism of free noradrenaline and, therefore, cause false-positive results.15 Eisenhofer et al.6 found that tricyclic antidepressants and phenoxybenzamine accounted for 44–45% of false-positive elevation in plasma and urinary noradrenaline.

When we applied criterion 2 (plasma noradrenaline plus adrenaline >2.96 nmol l–1 at 3 h), there were two false-positive results. In the first patient, there was no apparent explanation for this result. He was a 46-year-old male with a 2-year history of hypertension poorly controlled on three antihypertensive agents. Chest X-ray showed a right hilar mass, which turned out to be sarcoid. Computed tomography was negative for an adrenal mass and after 10 year follow-up, phaeochromocytoma has not been diagnosed. The second patient with a false-positive result was a 58-year-old male with type II diabetes, hypertension and a long-standing history of sweating. He had continued to take reboxetine, a noradrenaline reuptake inhibitor, before and during the clonidine suppression test. He had not informed us of this at any stage. Abdominal computed tomography showed no evidence of phaeochromocytoma. He is under continued review at the diabetic clinic, and blood pressure has been well controlled on three antihypertensive agents.

It is important to note that the type of catecholamine secreted by the tumour can affect the accuracy of the clonidine test. Tumours, which are predominantly adrenaline secreting may give rise to false-negative results if the diagnosis is based solely on the level of plasma noradrenaline. This is confirmed in our series. When Bravo's refined criterion using both adrenaline and noradrenaline is applied, the sensitivity increases. Just one patient had a predominantly adrenaline-secreting tumour, which may have been missed had adrenaline not been measured during the test. Likewise, those patients who have exceptionally rare tumours that secrete dopamine only, most of which are malignant, will not benefit from clonidine suppression testing.16 Only significant elevations of plasma dopamine will identify these tumours. Tumours, which are not predominantly noradrenaline secreting are more common in those with an inherited genetic basis. In our series, one patient had MEN-2B and another had neurofibromatosis. In both of these cases, tumours were predominantly noradrenaline secreting.

False-negative results may theoretically occur if the tumour has intermittent secretory activity, as measurement of the plasma catecholamines at the various time points may refer to different states of biochemical activity.8 Our series has shown this as not to be a major problem. If it does become such, the problem may be overcome by measuring levels of catecholamine metabolites, the metanephrines. These are constantly produced by the actions of catechol-O-methyltransferase on catecholamines leaking from stores within tumour cells and, therefore, show more consistent increases above normal in patients with phaeochromocytoma than plasma catecholamines, making this the more reliable test in these circumstances.17, 18, 19, 20 There is, however, still the possibility of false-positive results in which levels of metabolites are above the upper limit of normal, but not high enough to definitely diagnose phaeochromocytoma. Eisenhofer et al.6 have shown that, in such cases, clonidine suppression testing combined with measurement of plasma metanephrines can distinguish true- from false-positive results. In a series of 48 patients with phaeochromocytoma who underwent clonidine testing, 16 had normal levels or decreases in noradrenaline after clonidine. In contrast, plasma norepinephrine levels remained elevated in all, but two patients. Further studies using metanephrines during clonidine testing are required and this is a step, which we ourselves are undertaking. Currently, we are comparing, in a new series, the sensitivity and specificity of measurement of basal and 3 h post-clonidine plasma catecholamines and plasma metanephrines. This study should provide novel data on which is the more appropriate test.

Grossman et al.13 have assessed the clinical utility of combining clonidine suppression tests and glucagon provocation testing. When both glucagon and clonidine tests are negative, a diagnosis of phaeochromocytoma is highly unlikely. Glucagon tests must, however, be performed in carefully selected patients because of its potentially dangerous effects.21, 22 In the Grossman series, clonidine suppression test had a high sensitivity (97%), but low specificity (67%), whereas glucagon testing had low sensitivity (81%), but high specifity (100%).13 It may be that the low specificity of the clonidine test in their series was related to performance of the test on patients with lower resting catecholamine levels, and their important study may be of greatest relevance in the screening of those patients with gene mutations associated with phaeochromocytomas. Often these tumours are relatively quiescent.

In our analysis of the three literature-based criteria for a positive clonidine suppression test (Table 2), we have clearly shown that using criterion 2 (plasma noradrenaline plus adrenaline >2.96 nmol l–1 at 3 h post-clonidine administration or a baseline plasma noradrenaline plus adrenaline >11.82 nmol l–1) gave the best diagnostic accuracy with a sensitivity of 93% and a specificity of 95%. The diagnostic accuracy using this criterion would be improved further had the patient on antidepressant medication been removed from the analysis yielding a sensitivity of 100% and specificity of 98%.

In summary, in our regional endocrine centre, we have found that under carefully controlled conditions and using criteria based on the absolute level of noradrenaline plus adrenaline 3 h after administration of the drug as per Bravo et al., the clonidine suppression test was both safe and accurate in the investigation of patients referred to us with clinically suspected phaeochromocytoma and/or with abnormal urinary catecholamines. Further studies of the clonidine test using measurements of plasma metabolites are required and comparison of these with free catecholamines would be valuable.

References

  1. 1

    Ilias I, Pacak K . Current approaches and recommended algorithm for the diagnostic localization of phaeochromocytoma. J Clin Endocrinol Metab 2004; 89: 479–491.

    CAS  Article  Google Scholar 

  2. 2

    Manger WM . In search of pheochromocytomas. J Clin Endocrinol Metab 2003; 88: 4080–4082.

    Article  Google Scholar 

  3. 3

    Manger WM, Gifford Jr RW . Clinical and Experimental Pheochromocytoma. Blackwell Science: Cambridge, MA, 1996.

    Google Scholar 

  4. 4

    Manger WM, Gifford Jr RW . Phaeochromocytoma. J Clin Hypertens 2002; 4: 62–72.

    Article  Google Scholar 

  5. 5

    Lenders JW, Eisenhofer G, Mannelli I, Pacak K . Phaeochromocytoma. Lancet 2005; 366: 665–675.

    Article  Google Scholar 

  6. 6

    Eisenhofer G, Goldstein DS, Walther MM, Friberg P, Lenders JW, Keiser HR et al. Biochemical diagnosis of pheochromocytoma: how to distinguish true- from false-positive test results. J Clin Endocrinol Metab 2003; 88: 2656–2666.

    CAS  Article  Google Scholar 

  7. 7

    Lenz T, Ross A, Schumm-Draeger P, Schulte KL, Geiger H . Clonidine suppression test revisited. Blood Press 1998; 7: 153–159.

    CAS  Article  Google Scholar 

  8. 8

    Bravo EL, Tarazi RC, Fouad FM, Vidt DG, Gifford Jr RW . Clonidine-suppression test: a useful aid in the diagnosis of pheochromocytoma. N Engl J Med 1981; 305: 623–626.

    CAS  Article  Google Scholar 

  9. 9

    Bravo EL, Gifford RW . Phaeochromocytoma: diagnosis, localization and management. N Engl J Med 1984; 311: 1298–1303.

    CAS  Article  Google Scholar 

  10. 10

    Manger WM . Phaeochromocytoma: clinical experience and overview. Curr Probl Cancer 1985; 9: 1–89.

    CAS  PubMed  Google Scholar 

  11. 11

    Eisenhofer G, Goldstein DS, Stull R, Keiser HR, Sunderland T, Murphy DL et al. Simultaneous liquid- chromographic determination of 3,4-dihydroxyphenylglycol, catecholamines, and 3,4-dihydroxyphenylalanine in plasma, and their responses to inhibition of monoamine oxidase. Clin Chem 1986; 32: 2030–2033.

    CAS  PubMed  Google Scholar 

  12. 12

    Lenders JWM, Eisenhofer G, Armando I, Keiser HR, Goldstein DS, Kopin IJ . Determination of plasma metanephrines by liquid chromatography with electrochemical detection. Clin Chem 1993; 39: 97–103.

    CAS  PubMed  Google Scholar 

  13. 13

    Grossman E, Goldstein D, Hoffman A, Keiser HR . Glucagon and clonidine testing in the diagnosis of phaeochromocytoma. Hypertension 1991; 17: 733–741.

    CAS  Article  Google Scholar 

  14. 14

    Witteles RM, Kaplan EL, Roizen MF . Sensitivity of diagnostic and localization tests for phaeochromocytoma in clinical practice. Arch Intern Med 2000; 160: 2957–2963.

    Article  Google Scholar 

  15. 15

    Messerli FH, Frohlich ED . High blood pressure. A side effect of drugs, poisons, and food. Arch Intern Med 1979; 139: 682–687.

    CAS  Article  Google Scholar 

  16. 16

    Proye C, Fossati P, Fontaine P, Lefebre J, Decoulx M, Wemeau JL et al. Dopamine-secreting phaeochromocytoma: an unrecognised entity? Classification of phaeochromocytomas according to their type of secretion. Surgery 1986; 100: 1154–1161.

    CAS  PubMed  Google Scholar 

  17. 17

    Eisenhofer G, Walther M, Keiser HR, Lenders JWM, Friberg P, Pacak K . Volume plasma metanephrines: a novel and cost-effective test for pheochromocytoma. Braz J Med Biol Res 2000; 33 (10): 1157–1169.

    CAS  Article  Google Scholar 

  18. 18

    Goldstein DS, Eisenhofer G, Flynn JA, Wand G, Pacak K . Diagnosis and localization of phaeochromocytoma. Hypertension 2004; 43: 907–910.

    CAS  Article  Google Scholar 

  19. 19

    Sawka AM, Jaeschke R, Singh RJ, Young Jr WF . A comparison of biochemical tests for pheochromocytoma: measurement of fractionated plasma metanephrines compared with the combination of 24-h urinary metanephrines and catecholamines. J Clin Endocrinol Metab 2003; 88: 553–558.

    CAS  Article  Google Scholar 

  20. 20

    Eisenhofer G, Lenders JWM, Goldstein DS, Mannelli M, Csako G, Walther MM et al. Phaeochromocytoma catecholamine phenotypes and prediction of tumour size and location by use of plasma free metanephrines. Clin Chem 2005; 51 (4): 735–744.

    CAS  Article  Google Scholar 

  21. 21

    Mannelli M, Ianni L, Cilotti A, Conti A . Phaeochromocytoma in Italy: a multicentre retrospective study. Eur J Endocrinol 1999; 141 (6): 619–624.

    CAS  Article  Google Scholar 

  22. 22

    Elliot WJ, Murphy MB . Reduced specificity of the clonidine suppression test in patients with normal plasma catecholamines levels. Am J Med 1988; 84: 419–424.

    Article  Google Scholar 

Download references

Acknowledgements

We are grateful to Sister D Saunderson and the nursing staff of Ward 7D, Royal Victoria Hospital for their expertise. During the course of this analysis, Dr Claire McHenry, Research Fellow, was in receipt of funding from the R&D office, Northern Ireland.

Author information

Affiliations

Authors

Corresponding author

Correspondence to A B Atkinson.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

McHenry, C., Hunter, S., McCormick, M. et al. Evaluation of the clonidine suppression test in the diagnosis of phaeochromocytoma. J Hum Hypertens 25, 451–456 (2011). https://doi.org/10.1038/jhh.2010.78

Download citation

Keywords

  • clonidine suppression testing
  • diagnosis
  • phaeochromocytoma

Further reading

Search

Quick links