Clinical Evaluation of Urinary Cortisol Determinations by Competitive Protein-Binding Radioassay
BEVERLEY E. P. MURPHY,! with the Technical Assistance of Christa Jachan Clinical Investigation Unit, Queen Mary Veterans Hospital and Department of Investigative Medicine, McGill University, Montreal, Canada
ABSTRACT. Urinary cortisol determinations were carried out on samples from 162 subjects by a method described previously (J Clin Endocr 27: 973, 1967). After extracting the cortisol from 1 ml or less of a 24-hr urine volume into methy- lene chloride, it was measured according to its competition with tritiated cortisol for binding sites on the corticosteroid-binding globulin of human plasma. Values in healthy subjects (48 ± 32 ug/24 hr with a range of 0 to 108 ug/24 hr) did not differ significantly from those found in
obese, chronically ill and hypertensive patients. All 32 values found in 14 cases of Cushing’s syndrome exceeded 120 ug/24 hr. Subjects in whom adrenocortical secretion was suppressed by dexamethasone or who were known to have adrenocortical hypofunction had very low values (3 ±6 ug/24 hr). This determination can be carried out in 2-3 hr and is highly specific for cortisol. It is particularly useful clinically for the detection of adrenocortical hyperfunction. (J Clin Endocr 28: 343, 1968)
I TT WAS SHOWN by Cope and Black as long ago as 1959 (1) that urinary cortisol was a more reliable index of adrenocortical hyperfunction than 17-hydroxycorticoids (17-OHCS), 17-ketosteroids (17-KS), or 17-ketogenic steroids (17-KGS). Neverthe- less, clinical use of this determination has not become widespread as yet. There may be several reasons for this: 1) The deter- mination of urinary cortisol has hitherto been more complicated to carry out than are the other tests mentioned, 2) unless ACTH is given, it has little value in the diagnosis of hypofunction of the adrenal cortex, 3) there have been few corrobora- tive studies illustrating its usefulness as a routine procedure.
In our own laboratory plasma corticoid determinations by competitive protein- binding radioassay (2) have almost en- tirely replaced urinary determinations for the study of responses to ACTH, dexa- methasone (9a-fluoro-118,17,21-trihydroxy- 16a-methyl-41,4-pregnadiene-3,20-dione), metyrapone (Metopirone) and pyrogen. However, the urinary cortisol, as determined
by a similar method, has proved to be more valuable for the detection and evaluation of hyperfunction of the adrenal cortex. It is the purpose of the present paper to de- scribe the results obtained in our laboratory where this procedure has been used rou- tinely over a period of several years.
Materials and Methods
Cortisol-1,2-3H, specific activity 22.1 c/ mmole, was obtained from the New England Nuclear Corporation, 575 Albany St., Boston, Mass. 02118. On arrival it was diluted to a concentration of 10 uc/ml ethanol and stored at -10 C. If necessary, it was repurified by a single thin-layer chromatography.
Fuller’s earth, acid-washed, cat. 134-F, was obtained from British Drug Houses (Canada) Ltd., Montreal, and was used without further preparation.
Nonradioactive cortisol was obtained from Sigma Chemicals, 3800 DeKalb St., St. Louis 18, Mo. It was stored at -10 C in a concentra- tion of 10.0 ug/ml ethanol from which a working standard containing 0.1 ug/ml ethanol was made. The working standard was made up in small (10 ml) quantities since it was found that it deteriorated after several months if subjected to the frequent rapid changes of temperature occurring with daily use. CBG-isotope solution was made up as follows: 5.0 ml pooled human plasma, 0.4 ml of cortisol-3H solution and dis-
Received August 14, 1967; accepted October 25.
1 Scholar, Medical Research Council of Canada.
| Day | Room temperature | Refrigerated | Frozen | |||
|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | 4 | 4 | |
| Subject | ||||||
| 1 | 37 | 57 | 55 | 86 | 63 | 50 |
| 2 | 297 | 264 | 343 | 360 | 360 | 339 |
| 3 | 64 | 76 | 63 | 48 | 73 | |
| 4 | 23 | 10 | 46 | 70 | 24 | 18 |
| 5 | 52 | 23 | 52 | 33 | 55 | |
| 6 | - | 67 | 54 | 67 | 43 | 65 |
tilled water to 100 ml. (Note: the plasma is added after adding most of the water to avoid precipitating the protein with ethanol.)
Urine was collected in plastic or glass bottles for 24-hr periods. It was refrigerated until the initial determination was made, then frozen for storage, except for samples from pregnant women. These had been frozen for several months before the initial determination.
The principles and technical details of this urinary cortisol determination have been dis- cussed previously (3). Briefly, the procedure was carried out as follows:
A volume of 1/3000 of a 24-hr urine sample was pipetted into a small extraction flask and the volume made up to 3 ml with water. After shaking twice for 30 sec with 3 ml volumes of methylene chloride, the extracts were combined, and two 2 ml aliquots were transferred to small test tubes and evaporated to dryness. Alterna- tively, 1/10,000 of the 24-hr volume was ex- tracted in duplicate and evaporated to dryness. To provide standards, 0, 0.1, 0.2, 0.3 and 0.4 ml of a solution containing 0.1 ug/ml cortisol in ethanol was pipetted into small test tubes in duplicate and evaporated to dryness.
All the test tubes were then placed in a rack and 1.00 ml of CBG-isotope solution was added to each. After shaking for a few seconds, the rack was placed in a 45 C bath for 5 min, shaken again, then cooled to 10 C for at least 10 min. While at 10 C, 15 mg Fuller’s earth was added to each tube. The rack was shaken vigor- ously for 2 min and returned to the cold water bath for 10 min. After centrifuging for 2 min, 0.50 ml supernatant was transferred into vials containing 10 ml of counting solution. The vials were counted twice to a preset count of 4000 in an Ansitron liquid scintillation spectro- meter and the mean times of the standards in minutes were plotted against mug cortisol to provide a standard curve, as shown in Fig. 1. The amount of cortisol in the urine samples was read off according to the time required for counting and multiplied by 9 (or 10) to express the results as ug/24 hr.
When the cortisol content exceeded the range used (i.e., 360 or 400 µg/24 hr), 1/10th, or occasionally 1/100th, of the usual volume was processed as above. Samples giving results greater than 100 ug/24 hr were repeated and the mean of all 4 values recorded.
Results
1. Recovery of nonradioactive cortisol added to urine. Cortisol 30 mug was added to each of a series of test tubes and evaporated to dryness. An aliquot of a urine sample was added to each of two test tubes and carried through the procedure as described above. The results were compared to those of the same samples without added cortisol. Since § of the extract was added to the assay sys- tem, the difference was calculated as a % of 10 mug cortisol. The mean recovery was 96 ±7% SE for 18 urines.
2. Reproducibility. The standard deviation was calculated from values for 20 duplicate samples according to the formula
S = / 1
,
>(d)2 2 2n
where d =difference between duplicate values and n =number of duplicate pairs. Over the ranges of values 0-10 mug, 10-20 mug and 20-30 mug, the sD’s were ±1.2, ±1.7 and ±2.5 mug, respectively. These correspond to sD’s of ±11, ±15 and ±23 ug/24 hr over the ranges 0-90, 90-180 and 180-270 µg/24 hr, respectively. Since in practice all determinations were carried out in duplicate, the probable errors of recorded values over these ranges are ±8, ±11 and ±16 ug/24 hr, respectively.
| OBESE | AFEBRILE CHRON. | PROLONGED ACTH | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| HYPER- TENSIVE | ILL | LOW FEVER | ACUTE | CUSHING'S | 4-HRS | AD-CORT | |||||||||||
| HEALTH | AMBUL. | PREG. | (T≤100) | ILL | F | ACTH | SUPP. | HYPO-FN | |||||||||
| URINARY | 10,000 | F FF | |||||||||||||||
| CORTISOL | |||||||||||||||||
| ug/24 hours | ៛ | R | |||||||||||||||
| D 0 | R | PP P P P | |||||||||||||||
| c 0 | R | P P | |||||||||||||||
| 1,000 | 1 G H | ||||||||||||||||
| a | |||||||||||||||||
| 1 | .. | ||||||||||||||||
| Y 8 | |||||||||||||||||
| . | |||||||||||||||||
| 8 · 0010 | 7 | ||||||||||||||||
| 00 0 | |||||||||||||||||
| 000 8 | . | 10 | M | · | |||||||||||||
| 100 | 0 | · | .. | .. | |||||||||||||
| og | : | .. | 0000 | ||||||||||||||
| 00 | - | ||||||||||||||||
| A | .... | ... | |||||||||||||||
| 8 | .. | : | · | ||||||||||||||
| Tentative | .. · | ||||||||||||||||
| normal | 0 | . | |||||||||||||||
| range | 0 | ... | .. | ||||||||||||||
| o | VV | ||||||||||||||||
| · | ٥ | ... . | · | ||||||||||||||
| · | |||||||||||||||||
| 10 | o | . .. | · | B | W ٧ | ||||||||||||
| <10 | 0.0 | zzzz | YYYYYYX a | SS TT VV UUUUUU www | |||||||||||||
| MEANS: | 48 | 36 | 43 | 42 | 53 | 109 | 64 | 187 | 1740 | 2400 | 295 | 2 | 4 | ||||
| SD: | ±32 | ±21 | ±30 | ±23 | ±26 | ±54 | ±46 | ±104 | - | - | ±99 | t3 | ±8 | ||||
FIG. 1. Urinary cortisol in various conditions. Males are denoted by solid dots, females by open circles. Patients for whom more than one value is given are referred to by letters which correspond to those in Table 2.
3. Alteration of values on standing. Aliquots of six different urines were frozen, refrig- erated and left standing at room tempera- ture for four days. Results are shown in Table 1. In samples 1 and 4 the values in- creased after several days at room tempera- ture but otherwise the values were consis- tent with the calculated experimental error of the method.
4. Values in various conditions. The values were divided into 12 groups as shown in Fig. 1, where males and females are dis- tinguished by different symbols and indi- vidual patients are indicated by letters.
In cases where several specimens deter-
mined on the same patient were within the normal range, only the mean value for that patient is recorded. In cases where endo- crine abnormalities were known to be pres- ent, all the individual samples determined are recorded.
In the groups comprising healthy, chron- ically ill, obese, benign essential hyper- tensive subjects and those ambulatory sub- jects admitted for investigation, the values were consistently below 120 ug/24 hr and usually below 100 ug/24 hr. The mean for healthy females, 41 ±30 SD ug/24 hr, did not differ significantly from that for nor- male males, 56 ±33 SD ug/24 hr. Those for ambulatory males, 43 ±29 ug/24 hr, and
| Patient | Age | Sex | Diagnosis |
|---|---|---|---|
| A | 47 | M | Oat-cell Ca lung +Cushing's |
| B | 43 | M | Oat-cell Ca lung +Cushing's |
| C | 48 | M | Oat-cell Ca lung +Cushing's |
| D | 68 | M | Oat-cell Ca lung +Cushing's |
| E | 60 | M | Oat-cell Ca lung +Cushing's |
| F | 44 | F | Oat-cell Ca lung +Cushing's |
| G | 65 | F | Cushing's; ? Ca lung |
| H | 49 | F | Cushing's due to adrenal hyperplasia |
| I | 35 | F | Cushing's due to adrenal hyperplasia |
| J | 37 | F | Cushing's due to adrenal adenoma |
| K | 43 | M | Cushing's due to adrenal hyperplasia |
| L | 49 | F | Cushing's due to adrenal hyperplasia |
| M | 37 | F | Cushing's due to adrenal hyperplasia |
| N | 35 | F | Cushing's due to adrenal adenoma |
| P | 43 | M | Iatrogenic Cushing's; receiving ACTH 40 U daily |
| Q | 65 | M | Hypox. diabetic; given ACTH 25 U daily for 2 weeks |
| R | 62 | M | No endocrine disease; given ACTH 55 U daily for 2 weeks |
| S | 37 | F | Cushing's (M), post adrenalectomy; receiving dexamethasone 0.75 mg/day |
| T | 72 | M | Addison's disease |
| U | 75 | M | Bilateral adrenal hemorrhage |
| V | 65 | M | Hypox. diabetic (Q); receiving dexamethasone 0.5 mg/day |
| W | 43 | M | Addison's disease; receiving dexamethasone 0.5 mg/day |
| X | 46 | M | No endocrine disease; receiving dexamethasone 2 mg/day |
| Y | 45 | M | No endocrine disease; receiving dexamethasone 2 mg/day |
| Z | 23 | F | No endocrine disease; receiving dexamethasone 2 mg/day |
| ₡ | 19 | M | No endocrine disease; receiving dexamethasone 2 mg/day |
| B | 55 | M | No endocrine disease; receiving dexamethasone 2 mg/day |
| Y | 44 | M | Bronchogenic Ca; febrile |
| a | 75 | M | Ca larynx; bronchopneumonia; afebrile |
chronically ill males, 53 ±26, fell between those of the healthy groups. The means for obese males, 40 ±20 µg/24 hr, and obese females, 31 ±25, were both lower than those of healthy individuals of the same sex but the differences were not significant (p>0.05). Hypertensive individuals had a mean excretion of 42 ±24 ug/24 hr, entirely within the normal range.
Pregnant women had values which fre- quently exceeded the normal range. The mean, 109 ±53 ug/24 hr, was significantly different from those of the control groups (p <0.001).
In a group of mildly febrile patients (oral T 99-100 F) two values exceeded 120 ug/24 hr but the mean value of 64 ± 46 was not significantly higher than the control means (p>0.05). However, of 12 values obtained in five acutely ill subjects, four of whom were febrile, ten were above the nor- mal range and the mean value of 187 was significantly elevated (p<0.001), being about four times that of the control sub- jects.
In subjects receiving ACTH 75 U intra- venously over a period of four hours, the mean value was significantly increased, 295 ±99 (p<0.001), and with prolonged administration of long-acting ACTH every four hours, greatly elevated, viz., 2405 ug/24 hr.
Values in subjects with Cushing’s syn- drome all exceeded 120 ug/24 hr, those in individual patients being fairly consistent. The diagnoses in individual cases are given in Table 2. In those with Cushing’s syn- drome, these were made as follows: In cases A to G the diagnosis was based on urinary and plasma 17-OHCS, urinary 17- KGS and, in all but case G, the autopsy findings. In case G, the patient developed hypokalemic alkalosis and a Cushingoid appearance about three months after an abnormal shadow had been noted in a chest x-ray. There was no suppression with 8 mg dexamethasone/day. She deteriorated rap- idly and died; unfortunately, no autopsy was performed and the diagnosis of car- cinoma of the lung was not confirmed. In
cases H to N, the diagnosis was made on the basis of high basal plasma and urinary 17-OHCS, and/or high urinary 17-KGS and 17-KS, with lack of suppression of these values with 2 mg dexamethasone/day. Satisfactory suppression was achieved with 8 mg/day in all but J and N. The diagnosis was compatible with the findings at opera- tion in all but patient I, who was to have received pituitary irradiation but was lost to follow up. In some cases individual val- ues obtained for 17-OHCS, 17-KGS, 17-KS were equivocal and inconsistent but the final bulk of evidence was convincing in each case. Case P was a chronic severe asthmatic who had been injecting 20 to 60 U ACTH (Duracton) into himself each evening for 15 years. He had a typically Cushingoid appearance.
Subjects who had adrenocortical hypo- function, due either to Addison’s disease (T, U), adrenalectomy (S) or hypophy- sectomy (V), had significantly low values, the mean being 4.2 +7.7 ug/24 hr (p < .001). Those in whom adrenocortical func- tion was suppressed by dexamethasone administration (W to ß) had a similarly low mean value of 1.5 ±3.3 ug/24 hr.
Discussion
Several methods are currently available for the determination of urinary cortisol. Most require considerable purification and are therefore tedious and time-consuming for routine use. A single, rapid fluorescent technique which measures free 11-hydrox- ycorticosteroids was described by De Moor, Raskin and Steeno in 1960 (4). However, Espiner, in comparing this meth- od with one involving paper chromatog- raphy and incorporating an internal iso- topic standard to correct for losses, found that, although the correlation between the two measurements was very high, the per- centage contribution by true cortisol aver- aged only 33% and varied from 8 to 73% in different urine samples (5).
The method described here offers several advantages. It is very sensitive less than 1 ml of urine is required for duplicate de-
terminations-and thus small volumes of samples can be stored for repeat deter- minations if necessary. It is rapid: One technician can complete two urines and a control in duplicate in two hours and 15 minutes, including counting and calculat- ing, or can determine about 20 in duplicate in one day. No purification of reagents is required and no special glassware washing procedures need be employed. The specific- ity has been shown to be very high with respect to both steroidal and nonsteroidal substances (3). This high specificity is also confirmed by the fact that, with negligible losses of cortisol, the mean normal value was 48 ug/24 hr, a value which compares well with those in the literature, as quoted in Table 3.
Its accuracy in assessing adrenocortical function is attested to by the fact that ex- tremely low values were found in patients suppressed with dexamethasone (which itself has no effect on the assay), and in those known to have hypofunction of the adrenal cortex. Its use in detecting hypo- function is limited, however, by the fact that normal subjects occasionally have very low values also. On the other hand, hypo- function is much better studied by deter- mining the response to ACTH, metyrapone and pyrogen, etc., rather than by relying on basal values of metabolites which do not give any reliable idea of adrenocortical reserve and which may be frankly mislead- ing in cases of cirrhosis.
With respect to the detection of hyper- function, the results recorded here are in good agreement with those obtained by others (1, 8, 9, 11, 15, 18). Thus, the uri- nary cortisol would seem to be very reliable since there is no overlap of normal and abnormal values except in pregnancy. Un- like urinary metabolite (17-OHCS, 17- KGS) and cortisol secretion rate deter- minations, values in obesity are not ele- vated. Even if urine collections are not quite complete, there is little chance of missing an elevated value since the changes associated with hyperfunction are so large -e.g., in our experience and that of Cope
| Urinary cortisol (ug/24 hr) | Correction for recovery | Reference | |||
|---|---|---|---|---|---|
| Mean | Range | No. of subjects | |||
| 35 | 13-86 | 24 | yes | Ayres et al. | 1957 (6) |
| 44 | 0-120 | 120 | no | Cope & Black | 1959 (1) |
| 36 | 18-64 | 11 | yes | Jones et al. | 1959 (7) |
| 15 | 3-48 | 8 | no | Ross | 1960 (8) |
| 71 | 0-181 | 38 | yes | Rosner | 1963 (9) |
| 20 | 0-50 | 14 | no | Schteingart | 1963 (10) |
| 10 | 4-20 | 12 | no | Harris & Crane | 1964 (11) |
| 31 | 20-48 | 10 | no | Minick | 1964 (12) |
| 93 | 60-159 | 24 | no | Pal & Smith | 1965 (13) |
| 32 | 12-69 | - | - | Ertel & Peterson | 1966 (14) |
| 74 | 35-95 | 13 | yes | Espiner | 1966 (15) |
| 48 | 0-108 | 23 | negligible loss | Present study | |
| Mean | Range | No. of subjects | |||
|---|---|---|---|---|---|
| 191 | - | 419 | cortisol estimated to be 54% of total | De Moor et al. | 1962 (16) |
| 204 | 78-372 | 42 | - | Mattingly et al. | 1964 (17) |
| 259 | 109-577 | 36 | cortisol estimated to be 33% of total | Espiner | 1965 (5) |
and Black (1), the urinary cortisol rises approximately 8-fold compared with a 3- fold rise in 17-OHCS.
The extent of the hyperfunction in our cases of Cushing’s syndrome bore a definite relationship to their etiology. Thus, in this series, values in excess of 1000 µg/24 hr occurred only in cases where an ACTH- producing neoplasm was present.
Acknowledgments
The author is indebted to Dr. Anne Bradley, Dr. C. J. Pattee, Dr. Martin Hofmann and Dr. Leslie Kovacs for providing samples from many of the cases with Cushing’s syndrome from the Royal Victoria Hospital, and to Dr. C. J. Pattee, Director of the Clinical Investigation Unit, Queen Mary Veterans Hospital, for provision of laboratory fa- cilities.
References
1. Cope, C. L., and E. G. Black, Brit Med J 2: 117, 1959.
2. Murphy, B. E. P., In Bajusz, E. (ed.), An In- troduction to Clinical Neuroendocrinology, S. Karger, Basel/New York, 1967, p. 164.
3. J Clin Endocr 27: 973, 1967.
4. De Moor, P., M. Raskin, and O. Steeno, Ann Endocr (Paris) 21: 35, 1960.
5. Espiner, E. A., J Endocr 33: 233, 1965.
6. Ayres, P. J., O. Garrod, S. A. Simpson, and J. F. Tait, Biochem J 65: 639, 1957.
7. Jones, K. M., R. Lloyd-Jones, A. Riondel, J. F. Tait, R. D. Bulbrook, and F. C. Greenwood, Acta Endocr (Kobenhavn) 30: 321, 1959.
8. Ross, E. J., J Clin Endocr 20: 1360, 1960.
9. Rosner, J. M., J. J. Cos, E. G. Biglieri, S. Hane, and P. H. Forsham, J Clin Endocr 23: 820, 1963.
10. Schteingart, D. E., R. I. Gregerman, and J. W. Conn, Metabolism 12: 484, 1963.
11. Harris, J. J., and M. G. Crane, Metabolism 13: 45, 1964.
12. Minick, M. C., Metabolism 15: 359, 1966.
13. Pal, S. B., and V. K. Smith, Clin Chim Acta 12: 558, 1965.
14. Ertel, N. H., and R. E. Peterson, Excerpta Med, 2nd International Congress on Hormonal Steroids, May 1966, Milan, Italy (Abstract).
15. Espiner, E. A., J Endocr 35: 29, 1966.
16. De Moor, P., P. Osinski, R. Deckx, and O. Steeno, Clin Chim Acta 7: 475, 1962.
17. Mattingly, D., P. M. Dennis, J. Pearson, and C. L. Cope, Lancet 2: 1046, 1964.
18. Brooks, R. V., J. Dupre, A. N. Gogate, I. H. Mills, and F. T. G. Prunty, J Clin Endocr 23: 725, 1963.