Aldosterone-Producing Adrenocortical Carcinoma

Preoperative Recognition and Course in Three Cases

EUGENIO ARTEAGA, M.D .; EDWARD G. BIGLIERI, M.D .; CLAUDIO E. KATER, M.D .; JOSE M. LOPEZ, M.D .; and MORRIS SCHAMBELAN, M.D .; San Francisco, California

Three patients with primary aldosteronism due to adrenocortical carcinoma were studied, two with hyperaldosteronism alone and one also with hypercortisolism; in the later stages all three had hypersecretion of glucocorticoid and androgenic hormones. Although clinical presentations were similar to those of patients with benign adenoma, all had significantly higher concentrations of deoxycorticosterone and aldosterone and more profound hypokalemia. Stimulation with adrenocorticotropin in two patients showed a good cortisol response but no aldosterone response. The circadian rhythm for cortisol was normal but absent for aldosterone and deoxycorticosterone. Sequential 24-hour circadian studies in one patient showed that as the disease progressed, corticosterone and finally cortisol lost their circadian rhythms. Treatment with spironolactone, mitotane, or aminoglutethimide had transient clinical effects. The patients died 2 to 13 years later.

PRIMARY ALDOSTERONISM is usually caused by a benign adrenocortical adenoma (1, 2), occasionally by bilateral adrenocortical hyperplasia (3), and rarely by adrenal cancer (4). Ovarian tumors have also produced excessive amounts of aldosterone (5). Adrenocortical carcinomas that only produce excessive amounts of mineralocorti- coids such as aldosterone and deoxycorticosterone are few (4, 6, 7); they almost invariably also produce many other adrenocortical hormones. This situation occurred in the reported case of primary aldosteronism due to an adrenocortical carcinoma (8).

We studied three patients with primary aldosteronism due to adrenocortical cancer and identified unusual bio- chemical features that increase the index of suspicion for malignancy. The longitudinal assessment of basal steroid levels and the circadian rhythms of cortisol, corticoster- one, deoxycorticosterone, and aldosterone define the clin- ical course of this type of primary aldosteronism.

Case Reports

PATIENT 1

A 56-year-old white man sought medical attention for weak- ness and was found to have high blood pressure and hypokale- mia. Initial studies (Table 1; Figures 1 and 2) established a diagnosis of primary aldosteronism with hypokalemia and marked renin suppression. A large right adrenal tumor weigh- ing 18.6 g (3.5 × 2.4 × 2.6 cm) was removed during a laparo- tomy and malignancy established by capsular and blood vessel invasion by malignant cells. After a brief asymptomatic period with normal hormonal levels, spironolactone was given because of recurrent primary aldosteronism (Table 1; Figures 1 and 3).

From the Medical Service, San Francisco General Hospital Medical Center, and the Department of Medicine, University of California, San Francisco; San Fran- cisco, California.

Catherization of the left adrenal vein showed suppression of aldosterone concentration indicating the presence of function- ing metastases. A second laparotomy identified metastases in the liver and kidney, around the inferior vena cava. Reduction of tumor mass and treatment with spironolactone (up to 600 mg/d), aminoglutethimide, mitotane, cyclophosphamide, and fluorouracil failed to control blood pressure or potassium levels. Hypercortisolism and hyperaldosteronism progressed. The pa- tient’s initial Porter-Silber chromogen and 17-ketosteroid levels were 10.8 and 7.7 mg/d, respectively, and both eventually rose to more than 30 mg/d. The patient died 2 years after initial diagnosis.

PATIENT 2

A 47-year-old white woman sought medical attention for muscle weakness. Eight years previously, hypertension, hypo- kalemia, and an 8-cm left adrenal mass had been found. An adrenocortical carcinoma was diagnosed after the mass was re- moved. After surgery urinary aldosterone levels remained at 1.7 µg/24 h for 1 to 2 years. She remained free of symptoms for 8 years. A diagnosis of primary aldosteronism with hypokalemia and renin suppression was again made (Table 1; Figures 1 and 2).

Metastatic adrenocortical carcinoma compressing the left kidney at the hilum was found during laparotomy. Postopera- tive aldosterone production remained high and required spiro- nolactone for control of blood pressure and potassium levels. Two years later an epigastric mass appeared, with liver metasta- ses and compression of the inferior vena cava. Porter-Silber chromogen and 17-ketosteroid levels initially at 7.4 and 12.7 mg/d, respectively, had increased to 25.0 and 28.9 mg/d. Treat- ment consisted principally of spironolactone. Mitotane and am- inoglutethimide were ineffective and poorly tolerated. The pa- tient died 11 years after initial diagnosis.

PATIENT 3

A 32-year-old white woman was found to be hypertensive with weakness, headaches, and nocturia. Facial hair and acne had appeared. A diagnosis of primary aldosteronism (Table 1; Figures 1 and 2) was established. Renin levels were profoundly suppressed. A 246-g right adrenal mass was removed at laparo- tomy. Adrenocortical carcinoma was established. The contrala- teral gland was suppressed. Both cortisol and fludrocortisone were required for several months before full recovery in 13 months. Two and one-half years after surgery, symptoms and hypertension recurred as did primary aldosteronism. The pa- tient’s features were cushingoid. Liver and lung metastases were documented. Porter-Silber chromogen and 17-ketosteroid lev- els, initially 23.8 and 16.0 mg/d, rose to 44.9 and 18.6 mg/d. Administration of mitotane and spironolactone were ineffective. The patient died 8 months later, 3 years after initial diagnosis.

Materials and Methods

The patients were admitted to the Clinical Study Center at San Francisco General Hospital Medical Center. After equilib- ration was achieved on a constant metabolic diet containing approximately 120 meq of sodium and 70 meq of potassium, the following procedures were done after overnight recumbency.

Venous blood samples were obtained at 0800 h after over- night recumbency for measurement of plasma steroid and renin

Table 1. Features and Course in Three Patients with Primary Aldosteronism Due to Adrenocortical Carcinoma*
Chronology of TreatmentBlood PressurePlasma Potassium ConcentrationPlasma Aldosterone ConcentrationUrinary AldosteroneCircadian RhythmLocalization Technique
Aldosterone and DeoxycorticosteroneCortisol
Level Rhythm
LevelRhythm
mm Hgmeq/Lng/dlpg/d
Patient 1
First surgery
Preoperative200/1001.4217.053.01FNNIVP
(-); 1311-C
(-); CT
Early postoperative146/904.01.2-8.55.0NFNN(-) 1311-C (-)
7-12 mos postopera- tive194/962.452.080-520IFNN
Second surgery 6 mos postoperative195/952.2368-3765...IFIF
Patient 2
First surgery
Preoperative170/1002.0...IVP (+)
Second surgery
Preoperative180/1002.233.638IFNN1311-C
(-)
Early postoperative150/1003.049.876IFNN
6 mos postoperative160/1003.5243...
Spironolactone
3 years after second surgery190/1123.248.0352IFIF
Patient 3
Surgery
Preoperative200/1301.7-2.371.6157-200IFNFIVP
(+);Ad
ven (+)
Early postoperative130/903.85.0<1D...D
12 mos postoperative130/904.49.67NNN
2.5 yrs postoperative140/1102.319.8113IFIF

. IVP = intravenous pyelogram; 1111-C - 1)11-jodocholesterol, CT = computed tomography; Ad Ven = adrenal venography; N = normal; F = fixed; I = increased; D = decreased.

concentrations. Twenty-four-hour urine collections were ob- tained daily for sodium and potassium levels and for frequent measurements of Porter-Silber chromogen, 17-ketosteroid, and aldosterone levels.

After the 0800-h sample was obtained, the patients assumed the upright posture for 4 hours when a second sample was ob- tained.

The circadian rhythm was determined by venous sampling every 4 hours during a 48-hour period (10). After overnight recumbency, the patients continued to remain in the recumbent position for the next 24 h. On the second day, after the 0800-h sample was obtained, the patients remained upright until 2400 h (16 hours total), when they resumed a recumbent posture until 0800 h. In Patient 1 five 24-hour circadian studies were done with a protocol similar to that of the second day of the 48- hour test.

Venous plasma samples were obtained at 0800 h and 1 hour after an intravenous bolus of cosyntropin, 0.25 mg (Patient 1) or before and 2 hours after continuous adrenocorticotropin in- fusion (0.02 mg/kg-min; Patient 2).

Plasma and urinary sodium and potassium levels were mea- sured by standard internal flame photometry. Urinary excretion of Porter-Silber chromogens, used to assess 17-hydroxysteroid, and 17-ketosteroid levels were measured for 24 hours by stan- dard techniques, and aldosterone levels were determined by the radioimmunoassay of the chromatographically isolated aldos- terone 18-glucuronide after 24 hours of hydrolysis at pH 1 (11). Normal values in our laboratory are as follows: aldoster- one, 4 to 20 µg/24 h; Porter-Silber chromogens, 4 to 12 mg/24

h; and 17-ketosteroid, 8 to 20 mg/24 h (men), and 5 to 15 mg/ 24 h (women). Plasma concentrations of aldosterone, deoxy- corticosterone, 18-hydroxycorticosterone, and 18-hydroxyde- oxycorticosterone were measured by radioimmunoassay after derivatization and chromatographic isolation (12). Cortisol and corticosterone (13) were measured by competitive protein binding analysis after chromatographic isolation. Plasma renin concentration was determined by the radioimmunoassay of an- giotensin I generated in excess of sheep substrate (14). Normal values for each of these determinations are shown in Figure 2. Statistical analysis was done by unpaired Student’s t-test; p < 0.05 was considered significant.

Results

BASELINE DETERMINATIONS

Figure 2 shows the basal plasma values obtained dur- ing the first hospitalization in our 3 patients compared with levels of 26 patients with surgically confirmed aldos- terone-producing adenomas. All patients with carcinoma had an increase in precursors 18-hydroxycorticosterone and deoxycorticosterone. The 18-hydroxydeoxycorticos- terone, cortisol, and corticosterone levels, when com- pared with those of patients with aldosterone-producing adenomas, had greater mean values of aldosterone (p < 0.05) and deoxycorticosterone (p < 0.001) and hypokalemia was more profound (p < 0.01). However,

Figure 1. Circadian variation of aldosterone and cortisol in Patients 1 (square), 2 (circle), and 3 (triangle) with adrenal carcinoma. The values are percentage variations from the mean. The stippled area represents the normal circadian variation in five patients with an aldosterone-producing adenoma.

180

ALDOSTERONE % CHANGE

160

140

120

100

80

60

40

180

160

CORTISOL % CHANGE

140

120

100

80

60

40

UPRIGHT

08

16

24

08

16

24

08

HOURS

there is considerable overlap. Plasma aldosterone concen- trations did not change significantly in response to the upright posture (217 to 217 ng/dL, 40 to 45 ng/dl, and 66 to 72 ng/dL).

CIRCADIAN VARIATION

At the time of their initial studies all three patients showed absence of circadian variation of aldosterone and deoxycorticosterone regardless of posture (Figures 1 and 3). In Patients 1 and 2, the circadian rhythm variation of cortisol was normal. In Patient 3 cortisol rhythm was inconsistent and abnormal. Figure 3 shows a series of four circadian studies done in Patient 1 during different stages of the disease. Initially, aldosterone (217 µg/dL) and deoxycorticosterone (50.4 ng/dL) levels were high and without rhythm whereas cortisol (9.9 µg/dL) and

corticosterone (222 ng/dL) maintained a normal circadi- an rhythm. After surgery, cortisol and corticosterone levels and rhythm remained normal, whereas the consid- erably lowered aldosterone (8.5 ng/dL) and deoxycorti- costerone (7.4 ng/dL) levels were still without a circadi- an rhythm. Hyperaldosteronism recurred 12 months after surgery. The steroid patterns at this time were simi- lar to the preoperative value. Finally, 18 months after surgery, aldosterone (368 ng/dL) and deoxycorticoster- one (386 ng/dL) levels continued to increase, corticos- terone rose to a mean 24-hour level of 1757 ng/dL, which was five times higher than the mean value found before surgery. Cortisol increased to 19.7 g/dL without the previously normal circadian rhythm. Administration of 1 mg of dexamethasone at 2400 h failed to suppress cortisol (0800 h, 11.4 µg/dL).

ADRENOCORTICOTROPIN

Administration of adrenocorticotropin (ACTH) dur- ing treatment with spironolactone treatment in two of the patients was associated with an increase in cortisol and corticosterone in both and in deoxycorticosterone in one. Aldosterone failed to increase, in contrast to the increases seen in patients with an aldosterone-producing adenoma during spironolactone treatment (Table 2). No change was seen in plasma potassium or renin concentration.

Discussion

Adrenocortical carcinoma producing hyperaldosteron- ism is a rare disorder (4). In our experience it represents less than 2% of all the patients with an aldosterone- producing tumor referred to our study center.

These three cases are representative of the clinical and biochemical variation in cases of adrenocortical carcino- ma (6). The clinical picture varied from the pure hyper- aldosteronism that fulfills the established diagnostic crite- ria (Patient 1), to hyperaldosteronism with subsequent biochemical evidence of glucocorticoid excess (Patient 2), to hyperaldosteronism with Cushing’s syndrome (Pa- tient 3). In Patient 1, the tumor initially produced aldos- terone and its precursors in excess, but cortisol was with- in normal limits. Patient 3 had concurrent hypersecretion of mineralocorticoid, glucocorticoid, and androgenic hor- mones when first seen. The different spectrum of hor- mone secretion seen in our patients shows the potential capacity of cancer cells to indiscriminately produce a var- ied spectrum of adrenal steroids.

The basal plasma levels of aldosterone and deoxycorti- costerone were significantly higher in the patients with cancer when compared to levels seen in patients with al- dosterone-producing adenomas. The importance of this basal determination in the differential diagnosis is limited because of the overlap between both groups (Figure 2). The high levels of deoxycorticosterone in the presence of normal levels of cortisol and corticosterone and normal or slightly elevated values of 18-hydroxydeoxycorticos- terone confirm that deoxycorticosterone was overpro- duced by the tumor cells. The elevated levels of aldoster- one failed to change with the assumption of the upright posture, another feature similar to patients with adeno- ma.

Figure 2. 0800-hour plasma ster- oid, potassium, and renin concentrations after overnight re- cumbency in 26 patients with an aldosterone-producing adenoma (open circle) and in three patients with adrenal carcinoma (triangle, square, closed circle). Stippled ar- eas represent normal ranges. A Student's unpaired t-test was done to compare both groups. DOC=deoxycorticosterone; 18- OHB=18-hydroxycorticosterone; 18-OH DOC=18-hydroxydeoxy- corticosterone; PRC =plasma renin concentration.

p< 0.01

250

P<0.05

450

0

10.0

5

150

ALDOSTERONE ng/dl

o

350

8.0

8

4

-

100

O

18-OHB ng/dl

250

0

PRC ng/ml/h

6.0

PLASMA K mEq/L

80

200

4.0

3

o

60

150

000

00

1.5

2

40

100

1.0

8

1

20

50

.

5

P<0.001

20

o

1,000

88

60

30

0

.

CORTISOL ug/dl

15

50

0

18-OH DOC ng/dl

CORTICOSTERONE ng/dl

750

DOC ng/dl

o

8

40

20

10

0

30

500

%

00

o

0

20

8

10

5

250

10

0

o

In patients with an aldosterone-producing adenoma, regardless of posture and while plasma renin concentra- tion is suppressed and potassium is stable, aldosterone shows a circadian variation similar to that of cortisol, essentially following the circadian rhythm of endogenous ACTH (10). Our patients had the unique absence of a circadian variation of aldosterone or deoxycorticosterone during the study periods (Figure 1).

The series of circadian studies done in Patient 1 (Fig- ure 3) are of special interest because they show the bio- chemical progression of the disease. Before surgery (Fig- ure 3A), and when the disease recurred 8 months after surgery (Figure 3C), the high levels of deoxycorticoster- one and aldosterone without circadian variation support a tumor origin of the hormones. Cortisol and corticoster- one, on the other hand, were within normal limits and

with circadian rhythm. In the postoperative period (Fig- ure 3B), aldosterone decreased to a subnormal value whereas cortisol and corticosterone remained within nor- mal limits. The last study (Figure 3D) was done while the patient was being treated with mitotane and spirono- lactone and showed a clear disarray of the previously normal circadian rhythm of cortisol and corticosterone with an increase in the 24-hour mean value of these hor- mones and a lack of suppression by dexamethasone. All of these changes suggested a tumor capable of producing glucocorticoids in excess.

The lack of circadian rhythm for aldosterone could be due to the absence of an aldosterone response to ACTH. Acute administration of ACTH normally induces a brisk rise in aldosterone in patients with an aldosterone- producing adenoma, but failed to stimulate aldosterone

Table 2. Effect of Adrenocorticotropin (ACTH) on Steroid Levels in Primary Aldosteronism due to Carcinoma and Adenoma During Spironolactone Treatment
Patients with CarcinomaSix Patients with Adenoma (Mean ± SE)
Patient 1Patient 2
ControlACTHControlACTHControlACTH
Aldosterone, ng/dL190.4165.453.642.440.1 ± 3.4109.1 ± 40.1
Cortisol, µg/dL10.830.417.627.711.6 ± 1.431.0 ± 2.4
Deoxycorticosterone, ng/dl66.4148.440.233.815.9 ± 2.9107.3 ± 16.4
Corticosterone, ng/dl47638665121112442 ± 1613871 ± 519
18-Hydroxycorticosterone, ng/dl22630415114595.5 ± 11.6364.2 ± 118.8
18-Hydroxydeoxycorticosterone, ng/dL12.0122.017.625.09.6 ± 3.1112.7 ± 12.8
Figure 3. Circadian studies done in Patient 1 in different stages of the disease: before surgery (A): 2, 12, and 18 months after unilat- eral adrenalectomy (B, C, and D respectively).

400

A.

B.

20

300

15

200

10

100

5

400

C.

D.

20

300

15

200

10

100

5

08

16

24

08

08

16

24

08

HOURS

ALDOSTERONE

ng/dl

DOC

ng/dl

CORTISOL

pg/dl

01-0

CORTICOSTERONE

ng/dl ×1004-4

production in our patients (Table 2). Thus, an elevated aldosterone level that fails to increase with ACTH could be helpful in identifying a malignancy.

The biochemical characterization of patients with this type of primary aldosteronism is crucial in projecting the course of the disorder. Patient 2 had only hyperaldoster- onism when the disease recurred 8 years after the first operation, and finally developed biochemical evidence of Cushing’s syndrome. The series of circadian studies done in Patients 1 and 2 showed that initially the hyperproduc- tion of steroids was of mineralocorticoids alone. As the disease advanced, autonomous secretion of cortisol by the tumor occurred. Eventually, all three patients had some degree of glucocorticoid hypersecretion with the hyperal- dosteronism. Hypersecretion of several unregulated hor- mones appears to be characteristic of the last stages of this disease, like other adrenal carcinomas (6, 7). There are some cases in which the tumor apparently secretes only aldosterone (4). However, steroid data were insuffi- cient and prolonged follow-up of the biochemical changes was not reported.

Treatment with conventional doses of spironolactone (200 to 600 mg/d) can achieve normal serum potassium levels without a short-term increase in urinary aldoster-

one excretion (16) and reduce or normalize blood pres- sure as in patients with an aldosterone-producing adeno- ma (4). However, in the later stage, the same dosage of spironolactone is unable to achieve the same results. The apparent failure of this drug can be the result of the ele- vated mineralocorticoid hormone levels that compete with spironolactone at the receptor site.

The disease is fatal, but the survival period differs as in our patients: Patient 2 lived 11 years after the first opera- tion, 8 years without evidence of recurrence, whereas Pa- tient 1 survived only 2 years after surgical removal of the tumor. Treatment with mitotane, used when the disease recurred, did not effect a significant change in the clinical course although biochemically it was associated with some decrease in the hormone levels.

The initial clinical biochemical characteristics are simi- lar in patients with an aldosterone-producing carcinoma and adenomas. The initial values of plasma mineralocor- ticoids and serum electrolytes overlap between both enti- ties. In the late stages aldosterone production can be ex- tremely high and hypersecretion of glucocorticoids can become evident. The lack of circadian variation in aldos- terone and its failure to respond to exogenous ACTH seems to be characteristic of a carcinoma but not of an aldosterone-producing adenoma.

ACKNOWLEDGMENTS: Grant support: in part by U.S. Public Health Service Research Grants HL-11046 from the National Heart, Lung, and Blood Institute, and AM-06415 from the National Institute of Arthritis, Metabolism, and Digestive Diseases. The studies were done in the General Clinical Research Center at San Francisco General Hospital Medical Center (RR-00083) with support by the Division of Research Resources, National Institutes of Health. Dr. Arteaga is a Fogarty International Fellow, grant number F05TW03180-01. Dr. Kater is an established investigator of CNPq. Brazil.

Requests for reprints should be addressed to Edward G. Biglieri, M.D .; San Francisco General Hospital Medical Center, Room 321, Building 100, 1001 Potrero Avenue; San Francisco, CA 94110.

References

1. CONN JW. Primary aldosteronism: a new clinical syndrome. J Lab Clin Med. 1955;45:3-17.

2. FERRIS JB, BROWN JJ, FRASER R, LEVER AF, ROBERTSON JI. Primary aldosteronism. Clin Endocrinol Metab. 1981;10:419-52.

3. BIGLIERI EG, SCHAMBELAN M, SLATON PE, STOCKIGT JR. The inter- current hypertension of primary aldosteronism. Circ Res. 1970;27(suppl I):1-175-202.

4. SALASSA TM, WEEKS RE, NORTHCUTT RC, CARNEY JA. Primary al- dosteronism and malignant adrenocortical neoplasia. Trans Am Clin Climatol Assoc. 1975;86:163-72.

5. TODESCO S, TERRIBILE V, BORSATTI A, MANTERO F. Primary aldos- teronism due to a malignant ovarian tumor. J Clin Endocrinol Metab. 1975:41:809-919.

6. RICHIE JP, GITTES RF: Carcinoma of the adrenal cortex. Cancer. 1980;45:1957-64.

7. ALTERMAN SL, DOMINGUEZ C, LOPEZ-GOMEZ A, LIEBER AL. Pri- mary adrenocortical carcinoma causing aldosteronism. Cancer. 1969;24:602-9.

8. FOYE LV JR, FEICHTMEIR TV. Adrenal cortical carcinoma producing solely mineralocorticoid effect, Am J Med. 1955:19:966-75.

9. LUND JO, NIELSEN MD. Fludrocortisone suppression test in normal subjects, in patients with essential hypertension and in patients with various forms of aldosteronism. Acta Endocrinol. 1980;93:100-7.

10. SCHAMBELAN M, BRUST NL, CHANG BCF, SLATER KL, BIGLIERI EG. Circadian rhythm and effect of posture on plasma aldosterone concen- tration in primary aldosteronism. J Clin Endocrinol Metab. 1976:43:115-31.

11. BROWN RD, SWANDER A, MCKENZIE JK. Urine aldosterone radioim- munoassay: validation of a method without chromatography. J Clin En- docrinol Metab. 1976;42:894-930.

12. BIGLIERI EG, WAJCHENBERG BL, MALERBI DA, OKADA H, LEME CE, KATER CE. The zonal origins of the mineralocorticoid hormones in

the 21-hydroxylation deficiency of congenital adrenal hyperplasia. J Clin Endocrinol Metab, 1981;53:964-9.

13. MURPHY BE. Some studies of the protein-binding of steroids and their application to the routine micro and ultramicro measurement of various steroids in body fluids by competitive protein-binding radioassay. J Clin Endocrinol Metab, 1967;27:973.90.

14. STOCKIGT JR, COLLINS RD, BIGLIERI EG. Determination of plasma renin concentration by angiotensin I immunoassay: diagnostic import of precise measurement of subnormal renin in hyperaldosteronism. Circ

Res. 1975;29(suppl II):11175-191.

15. KEM DC, WEINBERGER MH, HIGGINS JR, KRAMER NJ, GOMEZ-SAN- CHEZ C, HOLLAND OB. Plasma aldosterone response to ACTH in pri- mary aldosteronism and in patients with low renin hypertension. J Clin Endocrinol Metab. 1978,46:552-60.

16. KATER CE, BIGLIERI EG, SCHAMBELAN M. ARTEAGA E. Studies of the impaired aldosterone response to spironolactone-induced renin and potassium elevations in adenomatous but not hyperplastic primary al- dosteronism. Hypertension. 1983;5:V115-21.