Further Studies of Adrenocortical Function in Patients with Carcinoma of the Lung*
EMILE E. WERK, JR., M.D., LEON J. SHOLITON, M.D. and RICHARD T. MARNELL, M.D. t Cincinnati, Ohio
I N the last ten years there has been increasing interest in patients who manifest Cushing’s syndrome in association with extra-adrenal cancer, particularly bronchogenic carcinoma. Prior to 1959 twenty such patients with broncho- genic carcinoma were mentioned in the litera- ture [1,2], one of whom presented features of a functioning carcinoid tumor [3]. Since then at least eighteen additional patients have been described [4-17], two of whom have been classified as having malignant carcinoid tumors [12,16] and two as unusual types of bronchial adenoma [11] .¿ Several clinical reviews dealing with this interesting association have been presented, pointing out that the lung carcinoma is usually of the oat or undifferentiated cell type [7,8,18,19]. Also, evidence has been offered recently that an ACTH-like material may be produced directly by the tumor [17,20].
In addition, information is at hand suggesting a more extensive role of adrenal cortical function in patients with bronchogenic carcinoma, the great majority of whom have no clinical evidence of Cushing’s syndrome. The observa- tion that bronchogenic carcinoma tends to metastasize to the adrenal glands has long been of interest. Sommers and his colleagues have demonstrated a high incidence of adrenocortical hyperplasia in patients with bronchogenic carcinoma [21-23]. Previous investigation of patients with bronchogenic carcinoma in our laboratory has demonstrated elevated levels of 17-hydroxycorticosteroids (17-OHCS) in
# We have not included the patients reported pri- marily to have thymoma in conjunction with Cushing’s syndrome although it has been suggested that such thymomas are closely related to oat cell bronchogenic tumors morphologically and do not resemble the usual varieties of thymoma [17].
plasma, § a tendency to increased urinary excretion of 17-OHCS and 17-desoxysteroids [1], a diminishing diurnal variation of steroid levels in plasma and urine with progression of the cancer [24] and a failure to account for the increased concentration of 17-OHCS in plasma on cmotional anxiety [25]. The increased incidence of adrenal hyperplasia has been con- firmed by autopsy. In addition, significantly widened adrenal cortices were found in patients with bronchogenic carcinoma whose glands fell short of the actual measurable criterion of adrenal hyperplasia [26].
The purpose of this report is to present a more extensive appraisal of certain aspects of this problem not apparent in the studies cited. The opportunity to do this was afforded principally by the collection of data from a larger group of patients, part of which had been included in the various studies from this laboratory already mentioned. Thus elaboration of the following main points regarding patients with broncho- genic carcinoma is provided in this presentation: (1) a correlation of adrenocortical function, including the response to ACTH administration, with patient survival time; (2) the relationship of brain metastases to adrenal hyperfunction; (3) a correlation of various steroid levels and fractions evaluating extra-adrenal steroid me- tabolism; and finally (4) the interrelationship of cancer cell type to these parameters.
CLINICAL MATERIAL AND METHODS
Patients with bronchogenic carcinoma were hos- pitalized in the Cincinnati Veterans Administration
§ Throughout this paper plasma 17-OHCS refers to the unconjugated component. Urinary 17-OHCS refers to the total of unconjugated and glucuronic acid con- jugated fractions unless otherwise indicated.
* From the Metabolism Section, Veterans Administration Hospital and the University of Cincinnati College of Medicine, Cincinnati, Ohio. Manuscript received April 27, 1962.
Hospital. Pertinent clinical information on many of these patients has been presented previously [1,24,25]. All had symptoms and varying clinical and x-ray evidence of their disease. A few patients were studied prior to exploratory thoracotomy. “Curative” lung resection was attempted in only three patients and subsequently proved to be unsuccessful in all three. Therefore the patients in this study represent those with progressive cancer. Of the sixty-seven patients with carcinoma of the lung the proportion of indi- vidual cell types was as follows: squamous twenty- three, undifferentiated (large cell) nineteen, oat (undifferentiated small cell) twelve, adenocarcinoma seven, and unknown six. The interval to death, from time of study and malignant cell type of respective patients, is depicted in the results.
There was no clinical evidence of Cushing’s syn- drome or mineralocorticoid excess in any of the patients described herein. The serum potassium and carbon dioxide levels, as well as those of other elec- trolytes, were normal in all at the time of study. Abnormalities in carbohydrate metabolism, namely, increased fasting blood glucose concentration and diabetic glucose tolerance curves, were observed as has been described in patients with various types of cancer [27,28]. Blood pressure was essentially normal in all patients. Hyperpigmentation of the skin was noticed in several, osteoarthropathy and gynecomastia in a few. Liver profile determinations (not including bromsulfalein tests) were normal in all but two terminal patients with jaundice. None of these scattered abnormalities were related to the steroid levels in the plasma or urine.
The control steroid levels in plasma and urine were obtained from hospitalized patients with chronic disorders of varying severity not neoplastic in nature. Clinical information on these patients also has been included in previous reports [1,24,25]. It should be noted that some of these patients had only mild elec- tive surgical conditions, whereas others were suffering from chronic debilitating medical illnesses associated with considerable weight loss. All were moderately well stabilized, not acutely ill at the time of study and ambulatory and alert. The age of the control group was comparable to that of the patients with cancer. For purposes of this study the data from these patients are included as one control group. In this group the heterogeneity did not influence values for 17-OHCS in plasma but it did affect values for the urinary excre- tion of steroids and creatinine since both of these values were higher in the “less sick” patients with less weight loss. As will be described, this heterogeneity of steroid levels in urine was corrected by relating steroid excretion to creatinine output.
Therapy with drugs known or suspected of inter- fering physiologically or chemically with adreno- cortical steroid determinations was discontinued in patients at least one week prior to collection of samples. Exceptions to this occurred in a few terminal
patients with cancer who were receiving Demerol® at the time of study.
Blood samples for the determination of 17-OHCS in plasma were drawn between 7 and 9 A.M. The urinary excretion of steroids and creatinine was determined from twenty-four-hour urine collections. When administered, ACTH was given intravenously, 25 units in 1 L. of normal saline solution or 5 per cent glucose in water, from 8 A.M. to 4 P.M. Blood samples were drawn before and at the end of the infusion. Urine collections were made during the twenty-four hours prior to ACTH administration (the baseline) and during the twenty-four hours encompassing the ACTH infusion.
The 17-OHCS in plasma and in urine were deter- mined by modifications of the methods of Nelson and Samuels [29] and Glenn and Nelson [30], respectively. Unconjugated 17-OHCS in urine were measured by extracting 50 to 200 ml. of urine, which was brought to pH 4.8 with acetate buffer, with 2.5 volumes of a chloroform-ether (1:3) mixture twice. The pooled extracts were divided into equal duplicates following solvent evaporation and then treated similarly to the procedure in the determination of total 17-OHCS in- cluding chromatography on Florisil® columns. The values for 17-OHCS in plasma take into account loss of steroid during Florisil chromatography. However, values reported for 17-OHCS in urine, both total and unconjugated, do not take into account loss of steroids at any step during the chemical assay. The average loss of steroid standards in urine has been found to be 15 per cent in the determination of total 17-OHCS (excluding beta glucuronidase hydrolysis) and 30 per cent for unconjugated 17-OHCS. Urinary levels of 17-ketosteroids (17-KS) were determined by the procedure of Drekter et al. [31] and 17-ketogenic steroids (17-KGS) by the direct method of Appelby et al. [32] as modified by Sobel et al. [33]. Urinary levels of creatinine were measured according to the technic of Folin and Wu [34]. Statistical analyses were performed according to standard methods [35].
Relationship of Steroid Excretion to Creatinine Output. In this study the urinary steroid values in the main are reported in terms of urinary output of creatinine. The principal reason for so doing is to correct for variations in lean body mass and age [36,37], which in turn appear to correlate with the excretion of creatinine [38,39]. A comprehensive evaluation of the relationship between urinary output of creatinine and steroid excretion has not been published to our knowledge, but the available evidence indicates a positive correlation of creatinine output with excre- tion of conjugated 17-OHCS [39,40] and cortisol secretory rate [40]. A steroid : creatinine ratio has been used as a “steroid index” to elucidate mild Cushing’s syndrome [41]. As part of a further study of this problem, we have found a significant coefficient of correlation between the urinary excretion of creati- nine and 17-OHCS and 17-KS in the forty control
63 PATIENTS
73 SAMPLES
1.75
1 .0.
1.50
CREATININE GRAMS PER 24 HOURS
CONTROL MEAN
1.25
1.00
1 S.D.
0.75
CANCER CELL TYPE
2 S.D.
☐
SQUAMOUS
0.50
☒ OAT
☒ UNDIFF.
ADENO.
☐ UNKNOWN
0.25
1.06’
1.18*
1.00°
0.84.
10.26
10.36
±0.31
+0.29
50
45
40
35
30
25
20
15
10
5
INTERVAL TO DEATH (WEEKS)
patients of this study, r being 0.62 and 0.69, respec- tively. Thus the application of steroid : creatinine ratios in the control group decreased the standard deviations of the means (Fig. 3 versus Fig. 4) and produced more homogeneous control reference data. In the group with carcinomas (sixty-one patients excluding repeat values), the corresponding coeffi- cients of correlation were 0.28 and 0.50, both of which were statistically significant, P < 0.05 and <0.01, respectively. The lower r values in the group with cancer compared to the control group were due to the wide scatter of values of those patients with cancer who survived five weeks or less and, in particular, to the high steroid levels in the patients with oat and undifferentiated cell types of cancer. When the levels from patients with cancer who survived less than five weeks were excluded, the coefficients of correla- tion between excretion of creatinine and 17-OHCS and 17-KS were 0.67 and 0.71, respectively (forty patients excluding repeat values). These analyses provide further support for a positive correlation between creatinine and steroid excretion and, as such, serve as justification for the use of the steroid : creatinine ratios in this study. Certain points, however, should be emphasized. As already implied, in the patients
with cancer studied within five weeks of death, no significant correlation was found between creatinine and steroid excretion. Furthermore the diminished creatinine excretion approximately five weeks before death (Fig. 1) brought about a rise in the steroid : cre- atinine ratios in this interval (and lowered them relatively in the longer survival periods). Although subject to question, still it would seem to us that these factors would not preclude the use of these ratios at the selective interval within five weeks of death in a relatively small number of patients since this rela- tionship correlated significantly at other survival intervals, implying its fundamental nature. None- theless, representative steroid data not based on creatinine excretion have been presented for com- parative purposes. (Fig. 4.)
RESULTS
The interval from the date of testing to the time of death was divided into the four periods (more than thirty weeks, thirty to fifteen weeks, fifteen to five weeks and five weeks or less) which coincided in part with changes in concentration of 17-OHCS in plasma as determined by
AMERICAN JOURNAL OF MEDICINE
72
55
67 PATIENTS
92 SAMPLES
45
CANCER CELL TYPI
☐ SQUAMOUS
☒ OAT
40
☒ UNDIFF.
☐ ADENO.
☐ UNKNOWN
35
30
☒
25
2 S.D.
1 S.D.
20
053
x
x
ox
X56 ☒
CONTROL MEAN
15
x
x
1 S.D.
x
10
16.1*
20.6°
19.1º
26.6°
+4.5
+6.0
+5.8
+8.2
50
45
40
35
30
25
20
15
10
5
INTERVAL TO DEATH (WEEKS)
PLASMA 17-OHCS ug./100 ml.
inspecting Figure 2. The same time intervals were then employed in arranging all the remain- ing data. None of the mean values for individual time periods and statistical analyses shown in the various figures and tables include any repeat values for the respective period, whereas repeat values were included in the scatter- grams. The term “elevated or high levels” refers to levels above the control mean value plus 2 standard deviations.
Relation of Steroid Levels to Survival Time and Cell Type of Cancer. An inverse correlation was found between the concentration of 17-OHCS in plasma and the interval to death. (Fig. 2.) Patients who lived for more than thirty weeks usually had normal levels in the morning. The one notable exception was a patient with brain metastases who subsequently was shown to have adrenal hyperplasia [1]. As survival time short- ened, the levels increased significantly above the control causing a plateau-like elevation between
six and thirty weeks before death, the mean for this entire period being 20 µg. per cent. Approxi- mately five weeks before death a further increase occurred (Fig. 2), the mean for the last five weeks being 26.6 ug. per cent.
There was no over-all correlation of plasma levels with cell type of cancer. (Fig. 2.) The highest values, however, were seen in patients with oat cell carcinoma.
Urinary excretion of 17-OHCS per gram creatinine, in relation to interval to death, is depicted in Figure 3. Prior to fifteen weeks before death the mean twenty-four-hour urinary level was 5.4 mg. per gm. creatinine, a value significantly above the control mean (P < 0.01). From about ten weeks to death, the mean levels increased further, with numerous abnormally elevated levels. The means for the fifteen to five week and the five week to death periods were each significantly higher than the control (P < 0.01), and the mean of the latter period
0
25
62 PATIENTS
84 SAMPLES
CANCER CELL TYPE
A
20
URINARY 17-OHCS mg. / Gm. CREATININE/ 24 HOURS
☐ SQUAMOUS
☐ OAT
x ☒
☒ UNDIFF.
☐ ADENO.
15
☐ UNKNOWN
10
2 S.D.
☒
☒
☒
☒
1 S.D.
☒
☒
☒
5
CONTROL MEAN
☒
☒
·
1 S.D.
00
5.4ª
5.3º
6.0°
8.6°
±1.5
±1.5
±3.7
±6.2
50
45
40
35
30
25
20
15
10
5
INTERVAL TO DEATH (WEEKS)
was significantly above the combined fifteen to thirty and greater than thirty weeks periods (P < 0.05). However, there was a considerable scatter of 17-OHCS excretion as death ap- proached (Fig. 3), ranging from high to low normal with the mean being elevated above those of the patients who survived the longest and the control group.
For comparison, the 17-OHCS levels in urine not related to creatinine are shown in Figure 4. The mean 17-OHCS levels for each time period were significantly elevated above the mean of the control group similar to the steroid : creati- nine ratios mentioned. But, in contrast to the latter ratios, there was generally a greater scatter of levels, the values being higher in patients who survived for more than fifteen weeks and lower in those who survived for five weeks or less. As mentioned previously, these last two differences appeared to be related to the diminishing creatinine excretion as death ap- proached. Thus, discounting creatinine output,
there was no further increase in the mean 17-OHCS levels within five weeks of death.
The relation of cancer cell type to 17-OHCS excretion can be determined by close inspection of Figures 3 and 4. Among the patients who survived for about ten weeks or more there were no differences in the 17-OHCS output among the patients with various types of carcinoma. Those with undifferentiated cell type of carci- noma had values similar to those with squamous cell and adenocarcinoma as well as the two patients with oat cell carcinoma. In contrast, within five to ten weeks of death the highest 17-OHCS values in urine were observed in patients with oat cell or undifferentiated cell carcinoma, although low to normal values also were found occasionally in patients with these two types of cancer. The mean urinary excre- tion of 17-OHCS per gram creatinine for those who survived five weeks or less was 12.5 mg. for those with oat and undifferentiated cell car- cinoma (n = 11) which was significantly higher
AMERICAN JOURNAL OF MEDICINE
o
65 PATIENTS
90 SAMPLES
CANCER CELL TYPI
☐ SQUAMOUS
☒ OAT
15
☒ UNDIFF.
☐ ADINO.
☐ UNKNOWN
URINARY 17-OHCS mg./24 HOURS
10
2 S.D.
X56
x
1 S.D.
xx
53
CONTROL MEAN
5
67
1 S.D.
6.1’
6.2’
5.6.
6.2°X
±3.0
±2.1
12.6
$3.0
50
45
40
35
30
25
20
15
10
5
INTERVAL TO DEATH (WEEKS)
(P < 0.02) than the mean of 5.7 mg. for those with squamous cell and adenocarcinoma (n = 10). Thus the higher over-all mean excretion of 17-OHCS, milligrams per gram creatinine, in the patients within five weeks of death (8.6 mg., Fig. 3) was accounted for by the high levels in the patients with oat and undifferentiated cell carcinoma. The patients with squamous cell and adenocarcinoma had a mean excretion of 5.7 mg. in this period, a value similar to the means for patients with all cell types of cancer in the longer survival periods and significantly elevated above the control (P < 0.01). It must be emphasized, however, that the high excretion of 17-OHCS in the patients with oat cell and undifferentiated cell carcinoma who survived less than five weeks was not due to the cell type per se but was also related to progression of the disease. Evidence for this is the fact that the excretion of 17-OHCS in two such patients was 3.5 and 4 mg. per gm. creatinine per twenty- four hours fifteen to thirty weeks prior to death and 14 and 23 mg. per gm. creatinine, respec- tively, during the five week survival period.
The urinary excretion of 17-KS, milligrams per gram creatinine (Fig. 5), tended to increase
within five weeks of death in the patients with cancer but the mean value (14.1) was not significantly greater than that of the patients in the longer survival periods or in the control group. The higher mean value could be accounted for again by patients with oat cell and undifferentiated cell carcinoma who had the highest excretion of 17-KS. In this five week period the mean 17-KS output, milligrams per gram creatinine, of the combined patients with oat cell and undifferentiated cell carcinoma (n = 9) was 18.7, which was significantly higher (P < 0.05) than the mean of 11.0 for the combined patients with squamous cell and adenocarcinoma group (n = 11). Similar to the excretion of 17-OHCS, the increased output of 17-KS in the former patients was also related to progression of the disease since two persons had had lower values of 17-KS in urine at an earlier time. In those who survived for longer periods there were no significant differences in the excretion of 17-KS between the patients with cancer and the control groups or among the patients with various cell types of cancer.
Forty-two patients, most of whom lived less than thirty weeks, were given ACTH. There
60 PATIENTS
BO SAMPLES
URINARY 17 - KETOSTEROIDS mg./Gm. CREATININE/24 HOURS
CANCER CELL TYPE
☐ SQUAMOUS
30
☐ OAT
☒ UNDIFF.
☐ ADENO.
25
☐ UNKNOWN
2 S.D.
20
04
1 S.D.
0000
☒
15
☒
☒
CONTROL MEAN
AX
☒
☒
00
10
1 5.0.
☒
oxo
5
10.4.
11.3*
11.2°
14.1º
12.0
±2.8
±3.5
17.0
50
45
40
35
30
25
20
15
10
5
INTERVAL TO DEATH (WEEKS)
42 PATIENTS
46 SAMPLES
CANCER CELL TYPE
120
☐ SQUAMOUS
☒ OAT
RISE (4) IN PLASMA 17-OHCS vg./100 ml.
☒ UNDIFF.
100
☐
ADENO.
☐
UNKNOWN
80
x
60
☐
2 S.D.
1 S.D.
CONTROL MEAN
x
☒
000
40
OA
1 S.D.
A
x
&
20
62.
38.6°
37.7.
37.9.
±13.6
±22.4
±13.8
±17.9
50
45
40
35
30
25
20
15
10
5
INTERVAL TO DEATH (WEEKS)
39 PATIENTS
42 SAMPLES
42.1
O 39.3
RISE ( 4 ) IN URINARY 17-OHCS mg./ Gm. CREATININE/24 HOURS
CANCER CELL TYPE
☐ SQUAMOUS
☒ OAT
☒ UNDIFP.
30
☐ ADENO.
☐ UNKNOWN
2 S.D.
20
1 S.D.
0
x
x
x
CONTROL MEAN
ox
X
x
x
x
10
1 S.D.
x
0
20.4*
16.3°
12.9.
9.4.
±4.5
±9.8
±7.0
±6.6K
50
45
40
35
30
25
20
15
10
5
INTERVAL TO DEATH (WEEKS)
appeared to be no significant change in the rise of plasma level of 17-OHCS in response to corticotrophin administration in these patients from thirty weeks to death (Fig. 6) and the mean rises were not different from that of the control group. There were a few low responses within five weeks of death. The results in those who survived for more than thirty weeks cannot be interpreted since there were only three patients in this group. One had an exaggerated response to ACTH, and the other two had high normal values.
The response of the urinary excretion of 17-OHCS per gram of creatinine to ACTH in the patients who survived for longer periods was higher than in the control group and diminished in the patients as death approached. (Fig. 7.) The increases in the patients who survived from thirty to fifteen weeks were significantly above those in the control group whereas in those who survived less than five weeks they were not. Again, one must consider with some reservation the responses in those who survived longer than
thirty weeks (only four patients); the mean rise for this group was above that of the control group as well as that of the patients with cancer who lived less than five weeks. The mean rise for the patients who survived for more than thirty weeks combined with those who survived from thirty to fifteen weeks was also significantly higher than that of the patients who survived less than five weeks.
The mean rise in the urinary excretion of 17-KS, milligrams per gram of creatinine, for the patients within five weeks of death was not significantly elevated above that in the control group (P > 0.05) whereas the mean rises in the other three periods were all significantly above that in the control group. (Fig. 8.) This sug- gested a tendency for a decreased response in 17-KS to ACTH in the patients within five weeks of death but it appeared much less prominent than the respective decrease in the urinary excretion of 17-OHCS.
No definite correlation was discernible be- tween cancer cell type and response to ACTH
RISE ( A) IN URINARY 17 - KETOSTEROIDS mg./Gm. CREATININE/24 HOURS
34 PATIENTS
39 SAMPLES
CANCER CILL TYPE
☐ SQUAMOUS
o
o
15
☒ OAT
☒ UNDIFF.
☒
☐ ADENO.
☐ UNKNOWN
10
2 S.D.
xo
☒
1 S.D.
O
5
CONTROL MEAN
x
1 S.D.
6.5°
6.9.
5.6°
6.1º
.
±1.8
±3.7
±4.1
±4.7
4-1.0
50
45
40
35
30
25
20
15
10
5
INTERVAL TO DEATH (WEEKS)
in either plasma or urinary steroid values. The rises in concentration of 17-OHCS in urine tended to be highest in the patients with squamous cell and adenocarcinoma and lowest in two persons with oat cell and undifferentiated cell carcinoma, respectively.
In Table I the values of the 17-OHCS are outlined separately from the standpoint of abnormally elevated levels to depict the instances of clearly demonstrable adrenal hyperfunction. Within five weeks of death, 62 per cent of the patients showed elevated plasma levels of 17-OHCS and 42 per cent increased urinary excretion of 17-OHCS; whereas in those who survived for more than thirty weeks, the inci- dence was 14 and 20 per cent, respectively. The abnormally high urinary excretion of steroids in the group five weeks prior to death was found predominantly in patients with oat cell and undifferentiated cell carcinoma as indicated previously. In contradistinction to the baseline levels, there was some suggestion that ACTH hyperresponsiveness was more apt to occur in patients who lived longer than thirty weeks, but
this may have been misleading due to the small number of patients in this category. Otherwise, survival time did not seem particularly to influence the incidence of hyperresponsiveness to the administration of ACTH. The over-all incidence of patients with an above normal increase in plasma levels was 14 per cent and in urinary excretion of 17-OHCS, milligrams per gram creatinine, 15 per cent, representing a combined total of 21 per cent, or nine of the forty-two patients tested. Seven of these nine patients had squamous cell and adenocarcinoma which was also consistent with the tendency for a higher incidence of hyperresponsiveness in the patients who survived longer.
In contradistinction, the concentration of 17-OHCS in plasma was not unusually low in any of these patients. Four patients with ad- vanced cancer, however, manifested low urinary excretion of 17-OHCS and/or low responsive- ness to ACTH administration as demonstrated in Table II. It was not possible to obtain urine collections in all instances, which reflects the severity of illness in these patients. Despite the
AMERICAN JOURNAL OF MEDICINE
| Analysis | Interval of Study to Death (weeks) | |||||||
|---|---|---|---|---|---|---|---|---|
| >30 | 30 to 15 | 15 to 5 | 5 or Less | |||||
| No. t | % | No. t | % | No. t | % | No. t | % | |
| Baseline levels | ||||||||
| Plasma 17-OHCS | 1/7 | 14 | 5/12 | 42 | 8/25 | 32 | 20/32 | 62 |
| Urinary 17-OHCS : creatinine | 2/10 | 20 | 1/12 | 8 | 6/23 | 26 | 11/26 | 42 |
| Rise in levels after ACTH | ||||||||
| Plasma 17-OHCS | 2/3 | 66 | 1/8 | 12 | 1/15 | 7 | 3/19 | 16 |
| Urinary 17-OHCS: creatinine | 3/4 | 75 | 2/11 | 18 | 1/16 | 6 | 1/11 | 9 |
* Elevated levels of 17-OHCS refers to those values above the control mean plus 2 standard deviations.
t No. refers to the number of patients with elevated levels in relation to the total number of patients studied in a given period. When there were multiple samples on a patient within one survival period only the initial determination was utilized for the compilation of this data.
| Patient | Cell Type | Interval to Death (wk.) | 17-OHCS in Plasma | 17-OHCS in Urine | |||
|---|---|---|---|---|---|---|---|
| Baseline (ug. %) | Rise after ACTH Administration (ug. %) | Baseline | Rise after ACTH Administration (mg./gm. Creatinine) | ||||
| mg./24 hr. | mg./gm. Creatinine | ||||||
| Ho | Squamous | 2 | 36 | 12 | |||
| Ki | Squamous | 7 | 10 | 1.6 | 2.5 | ||
| Ki (repeat) | Squamous | 3 | 26 | 22 | |||
| Ke | Undifferentiated | 2 | 16 | 18 | 1.5 | 1.0} | |
| Du | Oat | 1 | 49 | 33 | 9.4 | 26.8 | 1.5 |
* “Abnormally low” refers to values below the control range.
t Creatinine was not determined in the urine from this patient. However, we have included the rise in urinary 17-OHCS not based on creatinine since this was usually similar with adequate urine collections whether based on creatinine output or not.
low urinary levels in two patients (Ki and Ke) there was some increase in the concentration of 17-OHCS in plasma following ACTH adminis- tration. In addition, there was an increase in the urinary excretion of 17-KS of 6.9 mg. per twenty-four hours in one patient (Ke, not shown in Table II). In two patients (Ho and Du) low responses to ACTH administration in both plasma and urinary excretion were observed in association with elevated baseline
levels. Thus, none of these patients showed the usual generalized low pattern of adrenal insufficiency, but instead what appeared to be various types of poor adrenocortical reserve unrelated to cancer cell type.
Effects of Brain Metastases. This factor was considered separately to determine if central nervous system aberration, via hypothalamic- pituitary function, could bring about the abnor- mally elevated 17-OHCS levels. Fifteen patients
| Analysis | Baseline | Rise after ACTH Administration | ||
|---|---|---|---|---|
| No.t | % | No.t | % | |
| Plasma 17-OHCS | ||||
| Patients with brain involvement | 8/15 | 53 | 4/10 | 40 |
| Patients without brain involvement. | 20/41 | 49 | 0/24 | 0 |
| Urinary 17-OHCS: creatinine | ||||
| Patients with brain involvement | 8/14 | 57 | 1/6 | 17 |
| Patients without brain involvement. | 8/34 | 24 | 1/21 | 5 |
* All the patients with brain involvement were studied during the survival period fifteen weeks prior to death. One patient was also studied in a longer survival period (see text). Elevated levels of 17-OHCS refers to those values above the control mean plus 2 standard deviations.
t No. refers to the number of patients with elevated levels in relation to the total number of patients studied in a respective group.
(seven with oat cell and undifferentiated cell carcinoma, five with squamous cell and adeno- carcinoma, and three with unknown type of cancer) had clinical evidence of brain involve- ment. In Table III the incidence of high levels in these patients is compared to that in patients without clinical brain involvement during the fifteen week to death survival period. As can be seen, the incidence of high baseline values of 17-OHCS in plasma was similar in both groups. On the other hand, the incidence of abnormally increased baseline urinary excretion of 17-OHCS was greater in persons with brain involvement, however, many patients without obvious central nervous system dysfunction also had increased levels The higher incidence in the former group was not specifically related to cancer cell type. Since patients with advanced oat cell and undifferentiated cell carcinoma had higher urinary levels of 17-OHCS, however, the presence of brain involvement in such a patient almost invariably was associated with abnor- mally increased urinary excretion of 17-OHCS.
The few patients with hyperresponsiveness to ACTH administration in this survival period were found, for the most part, to have brain involvement regardless of cancer cell type.
Only one patient with brain involvement was studied during a longer survival period; all the 17-OHCS values were markedly increased, base- line and increases after ACTH administration.
No association of central nervous system aberration with a low 17-OHCS level and/or ACTH response was observed.
Correlation of Various Steroid Levels and Fractions (Extra-adrenal Cortisol Metabolism). In Figure 9 the plasma levels of 17-OHCS are plotted in relation to the urinary excretion of 17-OHCS in terms of creatinine output (excluding repeat values). Notwithstanding the scatter, there is a positive correlation (r = 0.56, P < 0.01). Fol- lowing the administration of ACTH the coefficient of correlation between rises in con- centration of 17-OHCS in plasma and increases in urinary excretion of 17-OHCS per gram of creatinine was 0.73 (n = 33, P < 0.01). There were occasional instances of lack of correlation between plasma and urinary responses following ACTH administration as shown previously in Table II.
There was a significant positive correlation between the baseline levels of 17-OHCS and 17-KS in urine, r being 0.70 (n = 61, P < 0.01). However, the correlation in rises of these respective levels after the administration of ACTH decreased considerably, r being 0.40 (n = 38, P = 0.01), and the difference between the two r values was significant (P < 0.05). This decrease was related apparently to the previously mentioned fact that the degree of rise in 17-KS levels in urine following ACTH ad- ministration did not diminish as death ap- proached as it did in 17-OHCS values. (Fig. 7 and 8.) None of these correlations were affected specifically by cell type of cancer.
Unconjugated 17-OHCS in Urine. Despite the presence of the general positive correlation between 17-OHCS concentration in plasma and urinary excretion of 17-OHCS per gram creatinine, disparity between the two was observed often. Increased plasma values from patients with all cell types of cancer were found in conjunction with normal levels of 17-OHCS in urine as demonstrated in the upper left rectangle of Figure 9. The reasons for this disparity could involve several aspects of extra- adrenal metabolism of cortisol. In an attempt to investigate certain of these aspects the un- conjugated 17-OHCS in urine were determined in twenty-one patients. In Figure 10 the results of these determinations are shown in terms of survival period and compared to the values for 17-OHCS in plasma and total 17-OHCS in urine, together with the percentage of uncon- jugated to total 17-OHCS in urine. The levels for unconjugated 17-OHCS in urine in the patients who survived less than five weeks were proportionately much higher than the total
AMERICAN JOURNAL OF MEDICINE
70
55
PATIENTS
35
PLASMA UNCONJUGATED 17 - OHCS ug./100 ml.
30
☒
☒
☒
☒
25
CONTROL MEAN + 2 S. D.
20
☒
CONTROL MEAN + 2 S. D.
r = 0.56
15
CANCER CELL TYPE
n = 55 P<0.01
☒
☐ SQUAMOUS
☐ OAT
10
☒
☒
☒
UNDIFF.
☐ ADENO.
☐ UNKNOWN
2
3
4
5
6
7
8
9
10
20
30
URINARY TOTAL 17 - OHCS mg. PER GRAM CREATININE PER 24 HOURS
17-OHCS in urine as compared to the values obtained in the patients who lived longer than thirty weeks. The mean twenty-four hour urinary excretion of unconjugated 17-OHCS was 216 µg. per gm. creatinine in the patients of the latter period and was increased approxi- mately fivefold (1,049 µg. per gm. creatinine) in the patients who survived five weeks or less, whereas the corresponding mean urinary excre- tion of total 17-OHCS was increased only about twofold, from 4.3 to 8.2. The mean plasma con- centration of 17-OHCS was increased from 13 to 33 ug. per cent in these respective periods. The enhanced rise in the urinary excretion of unconjugated steroids was reflected in the per- centage of unconjugated to total 17-OHCS in the urine. The mean percentage was significantly doubled in the patients who survived five weeks or less, irrespective of cancer cell type, as com- pared to the patients who survived for more than thirty weeks. The results from the patients who survived from five to thirty weeks were intermediate between those of the other two periods, as would be expected from the earlier data shown in this study.
metabolites, 17-KGS in urine was measured. The 17-KGS in urine include, normally, carbon 20 reduction products and in lesser amounts 6 beta hydroxy cortisol and pregnanetriol related compounds in addition to the “usual total” 17-OHCS. Thus the difference between the urinary values for 17-KGS and 17-OHCS will reflect the excretion of these other compounds. The 17-KGS level was determined in the urine of fifteen patients with cancer and twenty-five of the control subjects. The mean interval of testing to death in this group of patients with cancer was nineteen weeks; two patients lived less than five weeks; five patients five to fifteen weeks; six patients fifteen to thirty weeks; and two lived more than thirty weeks. Because the survival time was mainly intermediate in this small group it was considered as a whole for com- parison with the control group. It should be emphasized, however, that the concentration of 17-OHCS in plasma was significantly increased in these fifteen patients with cancer, the mean being 21 ug. per cent. Table IV summarizes a comparison of the mean levels of these urinary steroids between the two groups. It can be seen that the baseline levels of 17-KGS in urine varied between the two groups of patients com-
21 PATIENTS
20
5.0 ± 1.7
+
7.3 + 4.9
12.7 ± 4.0
PERCENT OF URINARY UNCONJUGATED TO TOTAL 17-OHCS
10
0
20
4.3 ± 1.4
5.0 + 3.1
0.2 ± 4.3
URINARY TOTAL 17-OHCS
mg./Gm. CREATININE PER 24 HOURS
10
0
2,000
216 + 85
440 + 441
1049 + 736
URINARY UNCONJUGATED
17-OHCS pg./Gm.
1,000
-.
CREATININE PER 24 HOURS
0
70
13 ± 4.2
19 ± 6.6
33 ± 18.4
PLASMA UNCONJUGATED
17-OHCS pg./100 ml.
35
CELL TYPE
*
0
D
UX
XOOK
O
O
0
A
0
60
50
40
30
25
20
15
10
5
4
3
2
1
INTERVAL TO DEATH (WEEKS)
* The designates for cell types are the same as in Figures 1 through 9.
t The figures represent the means + standard deviations for each respective interval and determi- nation. The differences of respective means between those who survived less than five weeks and in those who survived more than thirty weeks were all significant P < 0.05, <0.02, or ;< 0.01. The means for those who survived five to thirty weeks were not significantly different from those who were in the other two survival periods.
| Data * | Total 17-OHCS (mg./gm. Creatinine per 24 hr.) | 17-Ketogenic Steroids (mg./gm. Creatinine per 24 hr.) | 17-Ketogenic Minus 17-OHCS (mg./gm. Creatinine per 24 hr.) | Coefficient of Correlation of 17-OHCS and 17-Ketogenic | |
|---|---|---|---|---|---|
| r | P | ||||
| Baseline | |||||
| Carcinoma of lung (no. = 15) | 6.2 ± 3.3 | 19.9 ± 7.2 | 13.8 ±4.7 | 0.62 | <0.01 |
| Control (no. = 25) | 4.1 ±1.5 | 15.7 ±4.9 | 11.5 ±4.2 | 0.70 | <0.01 |
| P difference | <0.01 | <0.05 | >0.1 | ||
| Rise after ACTH Administration | |||||
| Carcinoma of lung (no. = 14) | +14.6 ± 9.6 | +30.7 ± 12.6 | +15.4 ± 5.2 | 0.88 | <0.01 |
| Control (no. = 24) | +10.4 ± 4.2 | +20.6 ± 10.4 | +10.2 ± 7.9 | 0.74 | <0.01 |
| P difference | >0.05 | <0.02 | <0.05 | ||
* Mean ± S.D.
parable to the 17-OHCS. Following ACTH administration the mean rise in the urinary excretion of 17-KGS, and the corresponding value for 17-KGS minus 17-OHCS, were both greater in the group with cancer than in the control group, whereas the respective rise in the urinary excretion of 17-OHCS was not quite statistically different between the two groups. A significant positive correlation was found be- tween the urinary excretion of 17-OHCS and 17-KGS in both groups before and after the administration of ACTH, as shown in Table IV.
COMMENTS
The data herein presented indicate a correla- tion in both baseline levels of 17-OHCS and responsiveness of 17-OHCS to ACTH adminis- tration with patients’ survival time. An inverse relationship was observed between the interval from time of study to death and concentration of 17-OHCS in plasma. A similar, but somewhat more complex, relationship was found with respect to urinary excretion of 17-OHCS per gram of creatinine. The urinary levels were elevated above the control levels in the patients with cancer who survived for the longest period (longer than thirty weeks) whereas the plasma values in the patients in this period were usually normal. In addition, although the mean value for urinary excretion of 17-OHCS per gram of creatinine increased further in patients within five weeks of death, there was considerable scatter of individual urinary levels including low values. Divergence in steroid excretion in patients with bronchogenic carcinoma was noted several years ago by Voigt and Kny [42] but was not apparent in our initial study because only a few patients who lived less than five weeks were included [1]. Moreover, a relationship of steroid excretion to survival time was not apparent in this latter report since the majority of the twenty-two patients were studied between five and thirty weeks prior to death. As the current study demonstrates, this period of survival appears to be intermediate in terms of steroid levels.
The mean response of 17-OHCS in plasma to corticotrophin stimulation was essentially nor- mal in patients who survived less than thirty weeks and did not change as the interval to death shortened. This finding in combination with the respective diminishing responses of 17-OHCS in urine from high to low normal suggests, according to the usual criteria, a
decline in adrenocortical reserve as survival time lessened. However, there were notable exceptions and a significant number of the over- all patients (nine of forty-two) demonstrated hyperresponsiveness to ACTH administration.
Thus, in evaluating adrenocortical function in patients with bronchogenic carcinoma, pro- gression of the disease must be taken into account. It should be emphasized that the patients in the present study did not include those with early or asymptomatic cases and hence the entire course of the disease has not been evaluated. It would seem that 17-OHCS levels in plasma would be normal in such patients but it would be of interest to ascertain if urinary excretion of 17-OHCS per gram of creatinine and responsiveness to ACTH stimula- tion in this type of patient might tend to be high in view of the results reported here in the “earliest” patients.
From a clinical view it would appear that the governing factor regulating the relationship between survival time and steroid levels may be the rate of change or lability of the disease process. Evidence for this was presented in a study of diurnal variation of adrenocortical function in which the normal diurnal variation of steroid levels in plasma and urine was found to be lacking in both patients with advanced bronchogenic carcinoma and acutely ill, con- fused patients without cancer [24]. It is also well known that 17-OHCS in plasma are elevated in diverse non-fatal clinical states in which lability of the process would appear to be a determinant.
The increased incidence of elevated steroid levels in patients with brain involvement in this study was probably in general related to the advanced stage of the disease rather than to the overt brain involvement per se. Since the criteria for central nervous system dysfunction were relatively crude, however, the possibility of localized brain damage causing hyperfunction in selective patients cannot be excluded. The original patient described with elevated steroid levels and adrenal hyperplasia (who is included in this study) had brain involvement, yet lived for over forty weeks, a survival time usually inconsistent with generalized advanced broncho- genic carcinoma [1]. It was suggested also that hyperresponsiveness to ACTH administration in the patients of this study with the most advanced disease was associated for the most part with brain involvement.
The finding of the considerably greater urinary excretion of steroids, both 17-OHCS and 17-KS, in patients with oat cell and un- differentiated cell carcinoma than in patients with squamous cell and adenocarcinoma within five weeks of death is of interest for several reasons. First, it is in keeping with the previously noted observation that it is the lability of the disease process which affects the degree of elevation in steroid levels. Undifferentiated and oat cell carcinoma of the lung are usually con- sidered the most rapidly progressing forms of lung cancer [43-45]. The rapidity in deteriora- tion of the patients with oat cell type is brought out in this study in that only a few such patients lived more than a few months. Secondly, the observations of the highest urinary excretion of steroids and the exceedingly high 17-OHCS levels in plasma in patients with oat cell car- cinoma agree with the observation of Williams and Sommers of a high incidence of Crooke’s cells, hyalinized basophils, in the pituitaries of such patients [23]. Crooke’s cells are generally considered to be indicative of increased circu- lating steroids of the cortisone type [46].
Finally, of particular significance, is the fact that the majority of reported cases of clinically manifest Cushing’s syndrome in association with bronchogenic carcinoma have been of the oat cell or undifferentiated cell type [7,8, 18,19]. Althouth patients with these types of carcinoma in this study had elevated levels without Cushing’s syndrome, some connection certainly is suggested. In the reported cases of Cushing’s syndrome with carcinoma of the lung the patients have usually had extraordinarily elevated steroid levels [19]. The steroid values in patients with oat cell and undifferentiated cell carcinoma in this study who did not exhibit Cushing’s syndrome were lower, but it must be emphasized that the degree of elevation of steroid levels in these latter patients was often comparable to that seen in patients with Cushing’s syndrome without extra-adrenal can- cer. It might be proposed that patients with progressive bronchogenic carcinoma may be relatively resistant to the development of the clinical manifestations of Cushing’s syndrome; hence, for its development extremely high steroid levels are required. Thus, patients with carcinoma of the lung with overt Cushing’s syndrome may represent just one extreme of the spectrum of adrenal hyperfunction in bronchogenic carcinoma.
In keeping with this proposal is the fact that elevated steroid levels were not entirely specific for patients with oat cell and undifferentiated cell carcinoma since they were found on occasion in patients with adenocarcinoma and squamous cell types. Marks et al. also have observed a patient with squamous cell carcinoma of the lung who had a high concentration of uncon- jugated and conjugated 17-OHCS in plasma, hyperresponsiveness to ACTH stimulation and adrenal hyperplasia [47]. Patients with squamous cell or adenocarcinoma who have high steroid levels may represent extremes for these par- ticular cell types or possibly may have a mixture of cell types including oat cell and undifferen- tiated cell, a situation which may exist fre- quently [45].
Although the complex problem of correlating in vivo steroid levels to postmortem adrenocortical findings is still under study, certain pertinent points might be mentioned. In general, patients who had elevated levels of 17-OHCS in urine and hyperresponsiveness to ACTH administra- tion tended to have adrenocortical hyperplasia also. But the presence of adrenocortical metas- tases at autopsy often prevented a clear appraisal of this relationship in individual cases. Two patients in whom 17-OHCS values in urine were elevated within one week of death, 10.0 and 11.8 mg. per twenty-four hours, respectively (unrelated to creatinine), had extensive adreno- cortical metastatic tumor growth with appar- ently marked destruction of normal cortical tissue, which indicates the potential error in interpreting the autopsy findings in such cases. On the other hand one patient (Ke, Table II), who was suspected antemortem of having real adrenocortical insufficiency, did have almost complete metastatic destruction of the adrenals. Patients with “poor adrenocortical reserve” did not necessarily demonstrate adrenal metastases suggesting that other unknown factors affected this lessened response to exogenous cortico- trophin stimulation.
To delve into a more comprehensive interpre- tation of the observations in this investigation, several aspects of steroid dynamics must be broached. A foremost question is whether the altered plasma and urinary steroid levels represent predominantly changes in adrenal steroidogenesis or in extra-adrenal steroid metabolism. In regard to the former the most definitive evidence elicited is the finding of increased urinary excretion of 17-OHCS. A
AMERICAN JOURNAL OF MEDICINE
decreased urinary excretion of 17-OHCS might be influenced by such extra-adrenal factors as plasma binding, hepatorenal dysfunction or possibly enhanced tissue steroid utilization. On the other hand increased excretion of 17-OHCS usually is thought to be indicative of augmented adrenal production of cortisol. The possibility of decrements in other metabolic pathways of cortisol resulting in only a relative increase in urinary excretion of 17-OHCS is unlikely in the present study in view of the positive correlations found between urinary levels of 17-OHCS, 17-KS and 17-KGS, respectively. The mean urinary level of 17-OHCS of patients with cancer was significantly elevated above the control level at each survival period, and within five weeks of death 42 per cent of the patients had abnormally increased excretion of 17-OHCS per gram of creatinine. Therefore, it seems that adrenal cortisol production in essence was enhanced in these patients. The fact that an over-all significant positive correlation was found between plasma and urinary levels of 17-OHCS suggests that the increased concentra- tion of 17-OHCS in plasma did reflect aug- mented adrenal cortisol secretion. For further confirmation of these points, cortisol secretion rates, using C14 cortisol, should be performed.
It is of interest that Nelson et al. have reported recently that “terminal patients” may have elevated levels of ACTH in plasma [48]. In addition to the increased baseline excretion of 17-OHCS, in the persons with cancer in the present study, further support for elevated circulating ACTH was the finding of a few patients who, within five weeks of death, showed hyperresponsiveness to exogenous ACTH ad- ministration as might be expected under condi- tions of chronic oversecretion of endogenous corticotrophin [49]. But it is difficult to fit into this formulation the finding of decreased responsiveness to exogenous ACTH administra- tion in the patients with cancer in general as the survival period decreased. It can be suggested only that other factors resulted in this apparently poor adrenocortical reserve including possibly extensive adrenal metastatic destruction in certain instances.
It is also obvious that in combination with the enhanced adrenal cortisol production there was considerable alteration in extra-adrenal cortisol metabolism in the patients with cancer as their disease progressed. High levels of 17-OHCS in plasma were often found in conjunction with
normal urinary excretion of total 17-OHCS. Urinary output of unconjugated 17-OHCS increased proportionately more than total 17-OHCS in the patients with most advanced disease, resulting in a higher ratio of uncon- jugated to total 17-OHCS. Shuster also has noted that the urinary excretion of unconjugated 17-OHCS was increased disproportionately in patients with bronchogenic carcinoma and certain other serious chronic illnesses [50]. Thus it appears that a progressive change, probably in the hepatic metabolism of cortisol, occurs as the cancer reaches the advanced stages. This interpretation is somewhat in conflict with that of an earlier study in which the rate of disappear- ance of exogenous cortisol from plasma was found to be the same in patients with carcinoma of the lung as in control patients [1]. The fact that the patients in the latter study were from “intermediate” survival periods might account at least in part for this disagreement. Further- more, the hepatic derangement described might represent the earlier stages of that demonstrated by Sandberg et al. in patients with miscellaneous diseases who die within two or three days [51].
An attractive hypothesis is the connection of the derangement noted by the high ratio of unconjugated to total 17-OHCS in urine to an hepatic action of increased circulating ACTH. The studies of Berliner et al. [52,53] and DeMoor et al. [54] indicate that ACTH inhibits the hepatic metabolism of cortisol; the former group of investigators suggest that this is due to inhibition of the rate of glucuronide conjugation, whereas the latter authors suspect an interfer- ence with ring A hydrogenation of the steroid. Interestingly the percentages of unconjugated 17-OHCS in urine found in the patients with cancer in this study, namely, 5 per cent for early cases and 12 per cent for advanced cases, are similar to those reported by Kornel and Hill in normal subjects before and after exogenous ACTH administration, 5 per cent and 10 per cent, respectively [55].
Another possibility is that both the increased 17-OHCS concentration of 17-OHCS in plasma and the high percentage of unconjugated 17-OHCS in urine in the patients with advanced cancer are due to an estrogenic-like effect mediated by enhanced plasma cortisol binding and reduced hepatic conjugation. Available information in this regard is conflicting. Sand- berg et al. have found increased plasma binding of C14 cortisol in women with advanced breast
cancer [56]. On the other hand, Farese and Plager have reported normal or decreased plasma protein binding of C14 cortisol in patients with various types of metastatic car- cinoma, including carcinoma of the lung, using an erythrocyte uptake technic [57].
Estrogen administration to subjects increases the ratio of unconjugated to total 17-OHCS in urine but this is chiefly due to a reduction in the output of the conjugated component of the total; the unconjugated fraction (chloroform extractable) has been reported to be unchanged [58,59], although some increase has been noted [60]. Consequently there is an over-all decrease in total 17-OHCS excretion and the cortisol secretory rate is diminished [58-61]. These estrogenic-induced changes in 17-OHCS in urine are not exactly similar to those observed in this study in the patients with advanced cancer; the net urinary excretion of uncon- jugated 17-OHCS was considerably increased and there was some increase in the excretion of total 17-OHCS. The elevation of the latter, however, was less than the former, indicating a relative decrease in conjugation (and/or reduction) of cortisol metabolites. Conceivably an estrogenic-like effect could be partly but not wholly responsible for the alterations noted. Further investigation is needed to elucidate this problem which is particularly appropriate in view of the estrogenic effects observed in some patients with bronchogenic carcinoma [23]. Moreover, urinary excretion of estrogen has been reported to be increased in patients with lung cancer when pulmonary osteoarthropathy is present [62].
The fact that the unconjugated 17-OHCS in urine actually increased in the patients with advanced lung cancer would tend to exclude diminished renal clearance as accounting in the main for the elevated concentrations of 17-OHCS in plasma.
Examples of other possible aberrations in extra-adrenal cortisol metabolism in the patients with cancer were the proportionately slightly higher increases in urinary levels of 17-KS and 17-KGS versus 17-OHCS following ACTH administration. This would suggest that cortisol was being metabolized relatively less into the 17-OHCS compounds and more into the 17-KS and 17-KGS particularly following the adminis- tration of ACTH and an increase in levels. Indeed the previously mentioned poor adreno- cortical reserve after ACTH administration in
patients with advanced cancer may have been chiefly the result of such a mechanism and thus may represent more apparent than real loss of adrenocortical reserve.
Further detailed study of various specific urinary metabolites is indicated to answer these questions satisfactorily. It seems feasible also that in the patients with advanced cancer the hepatic metabolism of cortisol might proceed at an increased rate via carbon 6 hydroxylation in view of the recent studies by Frantz et al. indicating that 6 beta hydroxycortisol may be the most plentiful unconjugated cortisol me- tabolite in human urine [63]. Since this polar compound is not extracted by chloroform-ether it was not included in the 17-OHCS assays in this study. Furthermore, an increase in urinary concentration of 6 beta hydroxy cortisol may not be reflected in the determination of 17-KGS [63]. It is of interest, also, that the excretion of 6 beta hydroxy cortisol is enhanced by estrogen administration [64].
Another important question concerns the source of the circulating corticotrophin in patients with cancer. Recently, Williams and Sommers found increased numbers of hyper- trophic amphophils in the pituitaries of patients who died from cancer, particularly broncho- genic carcinoma, which they regarded as indicative of a “pituitary-adrenocortical hyper- functional state” [23]. In contrast, however, Golden et al. noted in the adenohypophyses of thirteen patients who died of neoplastic disease “evidence of markedly increased secretory activity of acidophilic cells” which was inter- preted as representing enhanced growth hor- mone secretion in a response to malnutrition [65].
An alternate source of ACTH-like activity in patients with bronchogenic carcinoma may be the tumor itself. This possibility was mentioned by Thorne in 1952 in regard to patients with Cushing’s syndrome in conjunction with bron- chogenic carcinoma [66]. Support for this has been presented recently by Holub and Katz [20] and Meador et al. [17] who have demon- strated corticotrophic activity in the tumors of such patients with little or no ACTH activity in their pituitary glands. In any event, it seems premature at this time to extend the principal of the tumor per se producing ACTH to patients with bronchogenic carcinoma in general, i.e., those without Cushing’s syndrome. On the basis of current knowledge of pituitary-adrenal rela-
AMERICAN JOURNAL OF MEDICINE
tionships and on the findings of the wide spec- trum of adrenal hyperfunction in the patients in this study, it would seem that the ACTH in these patients most likely originated from their pituitary glands.
In conclusion,. consideration should be given to the crucial problem of the potential signifi- cance of adrenocortical function in terms of the growth of lung cancer. First, it must be stressed that it is not implied that such alteration in adrenocortical function as found in this study is restricted to patients with bronchogenic car- cinoma. As a matter of fact, this is clearly not the case [1,7,8,24,26]. From available evidence at this time it would appear that the degree of alteration in adrenocortical function is more pronounced in patients with lung cancer than with other types of cancer or chronic diseases in general. For this reason, and to maintain homogeneity, these studies have dealt mainly with patients with bronchogenic carcinoma. But do adrenal steroids have any basic connection with the support of malignant growth of various types? Is adrenal cortical hypersecretion a “non-specific” response to “stress” or is the hyperfunction related to secondary factors of normal cellular destruction, infection and mal- nutrition? In regard to this last point it is of interest that recently two types of adrenal cortical responses have been described in exogenous malnutrition depending upon the proportion of protein lack to total caloric insufficiency. Low urinary excretion of 17-OHCS was found in children with a dietary intake deficient in protein and a relative excess of calories, whereas increased urinary excretion of 17-OHCS and diminished response to ACTH administration was noted when the diet was insufficient in both protein and calories [67].
Obviously these questions cannot be answered with the present state of knowledge. On the basis of extensive study and review of tissue steroid metabolism, Berliner and Dougherty have suggested that the essential characteristics of a malignant cell may be related to its ability to metabolize cortisol [53]. The results of this investigation can be summarized only to indicate an integral position of cortisol secretion and metabolism in the metabolic alterations of progressive, invasive carcinoma in humans.
SUMMARY
Investigation of adrenocortical function in patients with progressive bronchogenic car- VOL. 34, FEBRUARY 1963
cinoma without Cushing’s syndrome has been extended with special reference to (1) disease progression (patient’s survival time); (2) cell type of carcinoma; (3) brain metastases; and (4) extra-adrenal cortisol metabolism.
The concentrations of 17-hydroxycorticos- teroids (17-OHCS) in plasma were usually normal in the patients who survived the longest, greater than thirty weeks; were increased ab- normally in patients who survived five to thirty weeks; and were further increased, often markedly, in patients within five weeks of death. The mean urinary excretion of 17-OHCS per gram creatinine was significantly higher than in the control group at all survival periods and increased further as death approached. Within five weeks of death, despite a considerable scatter of values, 42 per cent of the patients showed increased urinary excretion of 17-OHCS per gram creatinine, whereas in patients who survived for more than thirty weeks the incidence was 20 per cent. The abnormally high steroid excretion in the group five weeks before death was found predominantly in patients with oat cell and undifferentiated cell carcinoma; the mean urinary levels of both 17-OHCS and 17-ketosteroids (17-KS) were significantly higher in these patients than in patients with squamous cell and adenocarcinoma.
The mean response of 17-OHCS in plasma to ACTH stimulation was essentially normal and did not change as the interval to death shortened, but the response of 17-OHCS in urine dimin- ished, suggesting a decline in adrenocortical reserve. Nine of forty-two patients demonstrated hyperresponsiveness to ACTH administration, as reflected in 17-OHCS in plasma and/or urine.
An increased incidence of elevated steroid levels was found in patients believed to have brain metastases but, in general, it was probably related to the advanced stage of the disease rather than to specific brain involvement. In selective cases, however, brain metastases may have influenced adrenocortical hyperfunction.
Although the increased urinary excretion of 17-OHCS suggests augmented adrenal cortisol secretion, there was in addition considerable evidence for altered extra-adrenal cortisol metabolism in patients as the carcinoma progressed, as follows: High levels of 17-OHCS in plasma were found often in conjunction with normal urinary excretion of 17-OHCS. Urinary output of unconjugated 17-OHCS increased
proportionately more than total 17-OHCS in patients with the most advanced disease resulting in a higher ratio of unconjugated to total 17-OHCS. There was a proportionately slightly higher increase in urinary steroid levels of 17-KS and 17-ketogenic versus 17-OHCS follow- ing ACTH administration.
It is concluded that a spectrum of adreno- cortical hyperfunction and altered cortisol metabolism exists in patients with progressively invasive bronchogenic carcinoma as an integral part of the disease. The adrenocortical hyper- function is most pronounced in patients with advanced oat cell and undifferentiated carci- noma, indicating a connection between the high incidence of Cushing’s syndrome reported in these types of patients.
Acknowledgment: We would like to acknowledge the excellent technical assistance of Kenneth Allen, Margaret Ehrhart, Theresa Hughes, John Lindsley and Brandon Taylor. Dr. Joseph Incze, formerly in the Department of Pathology, Veterans Administration Hospital, kindly as- sisted in the pathology studies.
ADDENDUM
Since the completion of this manuscript we have observed that there is an increased pro- portion of urinary excretion of 6-hydroxy cortisol to total 17-hydroxycorticosteroids in patients with various types of advanced cancer; thus demonstrating further the altered metabo- lism of cortisol in these patients [68].
Attention should be called to the recently published article by Hymes and Doe which also pertains to adrenal function in cancer of the lung [69]. In general there is agreement between their study and that reported herein. Certain discrepancies appear to be related, at least in part, to differences in methodology.
REFERENCES
1. WERK, E. E., JR, and SHOLITON, L. J. Adreno- cortical function in carcinoma of the lung. Cancer, 13: 469, 1960.
2. CHENG, C. L., LIU, T. C., LIN, T. T. and LI, C. N. Cushing’s syndrome associated with bronchogenic carcinoma of the lung; report of a case with autopsy findings. Chinese M. J., 77: 260, 1958.
3. HARRISON, M. T., MONTGOMERY, D. A., RAMSEY, A. S., ROBERTSON, J. H. and WELBOURN, R. B. Cushing’s syndrome with carcinoma of bronchus and with features suggesting carcinoid tumour. Lancet, 1: 23, 1957.
4. MACPHEE, I. W. Cushing’s syndrome associated
with carcinoma of the bronchus. Brit. J. Surg., 46: 456, 1959.
5. GOLD, G. L., and SHNIDER, B. I. Some unusual syndromes associated with neoplastic disease. Ann. Int. Med., 51: 890, 1959.
6. LEWIS, J. M., SEARLE, N. B. and JORDAN, G. L. Cushing’s syndrome and bronchogenic carcinoma. Ann. Int. Med., 52: 1138, 1960.
7. ALLOTT, E. N. and SKELTON, M. O. Increased adrenocortical activity associated with malignant disease. Lancet, 2: 278, 1960.
8. BAGSHAWE, K. D. Hypokalaemia, carcinoma and Cushing’s syndrome. Lancet, 2: 284, 1960.
9. GELFMAN, N. A. Bronchogenic carcinoma with Cushing’s syndrome. A review with the report of an additional case. Am. Rev. Resp. Dis., 83: 555, 1961.
10. DOE, R. P., ZINNEMAN, H. H., FLINK, E. B. and ULSTROM, R. A. Significance of concentration of nonprotein-bound plasma cortisol in normal subjects, Cushing’s syndrome, pregnancy, and during estrogen therapy. J. Clin. Endocrinol., 20: 1484, 1960.
11. COHEN, R. B., TOLL, G. D. and CASTLEMAN, B. Bronchial adenomas in Cushing’s syndrome: their relation to thymomas and oat cell carcinomas associated with hyperadrenocorticism. Cancer, 13: 812, 1960.
12. CHRISTY, N. P. Adrenocorticotrophic activity in the plasma of patients with Cushing’s syndrome associated with pulmonary neoplasms. Lancet, 1: 85, 1961.
13. BILLINGHURST, B. M., THOULD, A. X., GALPIN, O. P. and HINTON, J. M. Carcinoma and Cush- ing’s syndrome. Brit. M. J., 2: 490, 1961.
14. VOGEL, M. D., KEATING, F. R., JR. and BAHN, R. C. Acute Cushing’s syndrome associated with bron- chogenic carcinoma. Proc. Staff Meet. Mayo Clin., 36: 387, 1961.
15. BORNSTEIN, P., NOLAN, J. P. and BERNANKE, D. Adrenocortical hyperfunction in association with anaplastic carcinoma of the respiratory tract. New England J. Med., 264: 363, 1961.
16. EscoVITZ, W. E. and REINGOLD, I. M. Functioning malignant bronchial carcinoid with Cushing’s syndrome and recurrent sinus arrest. Ann. Int. Med., 54: 1248, 1961.
17. MEADOR, C., LIDDLE, G. W., ISLAND, D., NICHOL- SON, W. E., LUCAS, C. P., NUCKTON, J. and LUETSCHER, J. A. Cause of Cushing’s syndrome in patients with tumors arising from “nonendocrine” tissue. J. Clin. Endocrinol., 22: 693, 1962.
18. KOVACH, R. D. and KYLE, L. H. Cushing’s syn- drome and bronchogenic carcinoma. Am. J. Med., 24: 981, 1958.
19. HILLS, A. G. and WOEBER, K. A. The syndrome of intrathoracic neoplasia with bilateral hyper- function of the adrenal cortex. Ann. Int. Med., 54: 1295, 1961.
20. HOLUB, D. A. and KATZ, F. H. A possible etiologic link between Cushing’s syndrome and visceral malignancy. Clin. Res., 9: 194, 1961.
21. PARKER, T. G. and SOMMERS, S. C. Adrenal cortical hyperplasia accompanying cancer. Arch. Surg., 72: 495, 1956.
22. SOMMERS, S. C. Host factors in fatal human lung cancer. Arch. Path., 65: 104, 1958.
AMERICAN JOURNAL OF MEDICINE
23. WILLIAMS, M. J. and SOMMERS, S. C. Endocrine and certain other changes in men with carcinoma of the lung. Cancer, 15: 109, 1962.
24. SHOLITON, L. J., WERK, E. E., JR. and MARNELL, R. T. Diurnal variation of adrenocortical function in non-endocrine disease states. Metabolism, 10: 632, 1961.
25. SHOLITON, L. J., WOHL, T. H. and WERK, E. E., JR. The relationship of two psychological variables, anxiety and hostility, with adrenocortical function in patients with lung cancer. Presented at the Veterans Administration Cooperative Chemo- therapy Studies in Psychiatry and Broad Research Approaches to Mental Illness, Cincinnati, Ohio, March 26, 1962.
26. SHOLITON, L. J., INCZE, J. S. and WERK, E. E., JR. Adrenocortical width in carcinoma of the lung. Cancer, 14: 105, 1961.
27. MARKS, P. A. and BISHOP, J. S. The glucose me- tabolism of patients with malignant disease and of normal subjects as studied by means of an intravenous glucose tolerance test. J. Clin. Invest., 36: 254, 1957.
28. MARKS, P. A. Disturbances in carbohydrate me- tabolism in cancer patients. In: Proceedings of The Endocrine Society. Springfield, Ill., 1961. Charles C Thomas.
29. NELSON, D. H. and SAMUELS, L. T. Method for determination of 17-hydroxycorticosteroids in blood; 17-hydroxycorticosterone in the peripheral circulation. J. Clin. Endocrinol., 12: 519, 1952.
30. GLENN, E. M. and NELSON, D. H. Chemical method for determination of 17-hydroxycorticosteroids and 17-ketosteroids in urine following hydrolysis with B-glucuronidase. J. Clin. Endocrinol., 13: 911, 1953.
31. DREKTER, I. J., HEISLER, A., SCISM, G. R., STERN, S., PEARSON, S. and McGAVACK, T. H. Deter- mination of urinary steroids. I. Preparation of pigment-free extracts and simplified procedure for estimation of total 17-ketosteroids. J. Clin. Endocrinol., 12: 55, 1952.
32. APPLEBY, J. I., GIBSON, G., NORYMBERSKI, J. K. and STUBBS, R. B. The use of sodium borohydride in the determination of 17-hydroxylated corti- costeroids. Biochem. J., 57: xiv, 1954.
33. SOBEL, C., GOLUB, O. J., HENRY, R. J., JACOBS, S. L. and BASU, G. K. Study of the Norymberski methods for determination of 17-ketogenic steroids (17-hydroxycorticosteroids) in urine. J. Clin. Endocrinol., 18: 208, 1958.
34. FOLIN, O. and Wu, H. System of blood analysis. J. Biol. Chem., 38: 81, 1919.
35. SNEDECOR, G. W. Statistical Methods; Applied to Experiments in Agriculture and Biology, 4th ed. Ames, Iowa, 1946. The Iowa State College Press.
36. PINCUS, J. In: Discussion of Sobel, H., and Marmor- ston, J. Hormonal influences upon connective tissue changes of aging. Recent Progress Hormone Res., 14: 457, 1958.
37. WEST, C. D., BROWN, H., SIMONS, E. L., CARTER, D. B., KUMAGAI, L. F. and ENGLERT, E., JR. Adrenocortical function and cortisol metabolism in old age. J. Clin. Endocrinol., 21: 1197, 1961.
38. PICÓN-REÁTEGUI, E. Creatinine excretion and body composition. Am. J. Clin. Nutrition, 10: 128, 1962.
39. ROMANOFF, L. P., RODRIGUEZ, R. M., SEELYE, J. M., VOL. 34, FEBRUARY 1963
PARENT, C. and PINCUS, G. The urinary excretion of tetrahydrocortisol, 3a-allotetrahydrocortisol and tetrahydrocortisone in young and elderly men and women. J. Clin. Endocrinol., 18: 1285, 1958.
40. ROMANOFF, L. P., MORRIS, C. W., WELCH, P., RODRIGUEZ, R. M. and PINCUS, G. The metabo- lism of cortisol-4-C14 in young and elderly men. I. Secretion rate of cortisol and daily excretion of tetrahydrocortisol, allotetrahydrocortisol, tetra- hydrocortisone and cortolone (20x and 208). J. Clin. Endocrinol., 21: 1413, 1961.
41. MELLINGER, R. C. and SMITH, R. W., JR. Studies of adrenal hyperfunction in 2 patients with atypical Cushing’s syndrome. J. Clin. Endocrinol., 16: 350, 1956.
42. VOIGT, K. D. and KNY, W. Untersuchungen zur Nebennierenrindenfunktion bei Bronchialcarci- nomen and Lungentuberkulosen. Arch. klin. Chir., 280: 74, 1954.
43. BOUCOT, K. R., HORIE, U. and SOKOLOFF, M. J. Survival in lung cancer. New England J. Med., 260: 742, 1959.
44. UMIKER, W. and FRENCH, A. J. Bronchogenic car- cinoma. A second look at cell type. Cancer, 13: 1053, 1960.
45. REID, J. D. and CARR, A. H. The validity and value of histological and cytological classifications of lung cancer. Cancer, 14: 673, 1961.
46. SOMMERS, S. C. Pituitary cell relations to body states. Lab. Invest., 8: 588, 1959.
47. MARKS, L. J., ANDERSON, A. E. and LIBERMAN, H. Carcinoma of the lung associated with marked adrenocortical hyperplasia and adrenal hyper- responsiveness to ACTH in the absence of Cush- ing’s syndrome. Ann. Int. Med., 54: 1243, 1961.
48. NELSON, D. H., COOPER, C. E., GOLD, E. M. and RUTHERFORD, E. R. Levels of plasma ACTH in surgical and other acutely ill patients. J. Clin. Invest., 40: 1067, 1961.
49. NUGENT, C. A., EIK-NES, K., KENT, H. S., SAMUELS, L. T. and TYLER, F. H. A possible explanation for Cushing’s syndrome associated with adrenal hyperplasia. J. Clin. Endocrinol., 20: 1259, 1960.
50. SHUSTER, S. Adrenal function in chronic wasting diseases. J. Clin. Endocrinol., 20: 675, 1960.
51. SANDBERG, A. A., EIK-NES, K., MIGEON, C. J. and SAMUELS, L. T. Metabolism of adrenal steroids in dying patients. J. Clin. Endocrinol., 16: 1001, 1956.
52. BERLINER, D. L., KELLER, N. and DOUGHERTY, T. F. Tissue retention of cortisol and metabolites induced by ACTH: an extraadrenal effect. Endocrinology, 68: 621, 1961.
53. BERLINER, D. L. and DOUGHERTY, T. F. Hepatic and extrahepatic regulation of corticosteroids. Pharmacol. Rev., 13: 329, 1961.
54. DEMOOR, P., HENDRIKX, A. and HINNEKENS, M. Extra-adrenal influence of corticotropin (ACTH) on cortisol metabolism. J. Clin. Endocrinol., 21: 106, 1961.
55. KORNEL, L. and HILL, S. R., JR. Paper chromato- graphic pattern of endogenous urinary cor- ticosteroids in normal subjects. Metabolism, 10: 18, 1961.
56. SANDBERG, A. A., SLAUNWHITE, W. R., JR. and CARTER, A. C. Transcortin: a corticosteroid-
binding protein of plasma. III. The effects of various steroids. J. Clin. Invest., 39: 1914, 1960.
57. FARESE, R. V. and PLAGER, J. E. The in vitro red blood cell uptake of C14-cortisol; studies of plasma protein binding of cortisol in normal and abnormal states. J. Clin. Invest., 41: 53, 1962.
58. CARTER, A. C., FELDMAN, E. B. and WALLACE, E. Z. The effects of steroids on the levels of the plasma 17-hydroxycorticosteroids and the serum protein-bound iodine. In: Biological Activities of Steroids in Relation to Cancer, p. 77. Edited by Pincus, G. and Vollmer, E. P. New York, 1960. Academic Press.
59. WALLACE, E. Z. and CARTER, A. C. Studies on the mechanism of the plasma 17-hydroxycorti- costeroid elevation induced in man by estrogens. J. Clin. Invest., 39: 601, 1960.
60. LAYNE, D. S., MEYER, C. J., VAISHWANAR, P. S. and PINCUS, G. The secretion and metabolism of cortisol and aldosterone in normal and in steroid- treated women. J. Clin. Endocrinol., 22: 107, 1962.
61. PETERSON, R. E., NOKES, G., CHEN, P. S., JR. and BLACK, R. L. Estrogens and adrenocortical func- tion in man. J. Clin. Endocrinol., 20: 495, 1960.
62. GINSBURG, J. and BROWN, J. B. Increased oestrogen excretion in hypertrophic pulmonary osteo- arthropathy. Lancet, 2: 1274, 1961.
63. FRANTZ, A. G., KATZ, F. H. and JAILER, J. W. 60-hydroxycortisol and other polar corticosteroids: measurement and significance in human urine. J. Clin. Endocrinol. 21: 1290, 1961.
64. KATZ, F. H., LIPMAN, M. M., FRANTZ, A. G. and JAILER, J. W. The physiologic significance of 66-hydroxycortisol in human corticoid metabo- lism. J. Clin. Endocrinol., 22: 71, 1962.
65. GOLDEN, A., BONDY, P. K., CHAMBERS, R. W. and GEIGER, C. L. Adenohypophyseal changes in patients dying of neoplastic disease. Yale J. Biol. & Med., 33: 299, 1961.
66. THORNE, M. G. Cushing’s syndrome associated with bronchial carcinoma; enquiry into relationship of this syndrome to neoplastic disease. Guy’s Hosp. Rep., 101: 251, 1952.
67. CASTELLANOS, H. and ARROYAVE, G. Role of the adrenal cortical system in the response of children to severe protein malnutrition. Am. J. Clin. Nutri- tion, 9: 186, 1961.
68. WERK, E. E., JR., SHOLITON, L. J., MACGEE, J. and MARNELL, R. T. Altered cortisol metabolism in advanced cancer. (Abstract.) Clin. Res., 10: 300, 1962.
69. HYMES, A. C. and DOE, R. P. Adrenal function in cancer of the lung with and without Cushing’s syndrome. Am. J. Med., 33: 398, 1962.