Pancreas Vol. 5, No.5 , pp. 515-518 0 1990 Raven Press, Ltd., New York
Pancreatic Cancer and Androgen Metabolism: High Androstenedione and Low Testosterone Serum Levels Carlos Ferniindez-del Castillo, Guillermo Robles-Diaz, "Vicente Diaz-SBnchez, and "Araceli Altamirano Pancreas Clinic and *Department of Reproductive Biology, Instituto Nacional de la Nutricih "Salvador Zubircin," Mexico City, Mexico
Summary: Serum androstenedione and testosterone levels were measured in 39 male patients with pancreatic cancer, and compared with the values obtained from 37 male patients with chronic pancreatitis or benign obstructive jaundice, and with those from 36 male patients with other gastrointestinal malignancies. Mean androstenedione values were significantly higher in the pancreatic cancer patients when compared to both control groups, and mean testosterone levels were significantly lower. The testosterone/androstenedione ratio was calculated and was also found to be significantly lower in the patients with pancreatic cancer. There was no difference in this ratio or in the androstenedione or testosterone levels when comparing both control groups. In two patients with stage I pancreatic cancer, serum androstenedione and testosterone levels were significantly altered, and returned to normal values after successful resection. These results confirm previous findings indicating that there is signifcant derangement in the androgen profile of patients with pancreatic cancer. Key Words: Pancreatic cancer-Androstenedione-TestosteroneAndrogens.
Cancer of the pancreas is a highly lethal disease whose frequency is increasing (1). Its high mortality is due to late diagnosis, and therefore many efforts are currently underway to detect the disease at earlier stages and to make a better differential diagnosis (2). Since pancreatic cancer affects more men than women (3,4), several clinical and experimental studies have focused on the possible role of androgens as indicators andor etiological factors in pancreatic carcinogenesis (5-9). Low serum testosterone levels have been found consistently in patients
with pancreatic carcinoma (5-7). Moreover, a low testosterone/dihydrotestosteroneratio has been described (7), suggesting an increased testosterone uptake by the tumor and increased transformation to the reduced metabolite, 5-a-dihydrotestosterone. In experimental models of pancreatic carcinogenesis, the tumor growth rate is stimulated by testosterone, whereas it is inhibited by antiandrogens or castration (8,9). These studies suggest that testosterone exerts atrophic effect on the pancreatic carcinoma. We have recently shown that the isolated per-
Address correspondence and reprint requests to Dr. G. Robles-Diaz at Department of Gastroenterology,Instituto Nacional de la Nutrici6n Sz, Vasco de Quiroga #is, nalpan 14000, ~ 6 ico D.F., Mexico.
hormones in male patients with cancer ~ and - in two Control groups (one with benign diseases of the pancreas or lower biliary tract, and the other
. ,
515
C . FERNANDEZ-DEL CASTILLO ET AL.
516
with gastrointestinal malignancies other than cancer of the pancreas).
PATIENTS AND METHODS Peripheral venous blood samples were obtained from 111 male patients who were divided into three groups. Group 1 comprised 39 patients with pancreatic carcinoma, aged 57 11 years (mean SD). The diagnosis of pancreatic cancer was histologically confirmed in 32 cases, and established by surgical exploration and compatible clinical course in the remaining 7 patients. According to AJCC staging (ll), seven patients were in stage I, two in stage 11, four in stage 111, and 26 in stage IV. Group I1 comprised 37 patients with benign pancreatobiliary disease, aged 54 & 14 years. Twentyeight had chronic pancreatitis, which was diagnosed on the basis of a compatible clinical picture and the presence of pancreatic calculi and/or abnormal exocrine function tests. The remaining nine patients had benign obstructive jaundice secondary to stones or bile duct stenosis. Group 111 comprised 36 patients with histologically proven gastrointestinal malignancies other than cancer of the pancreas, aged 58 ? 17 years. These carcinomas were distributed as follows: 15 gastric, 8 colorectal, 5 ampullary, 4 esophageal, 3 bile duct, and 1 hepatic. All blood samples were collected after an overnight fast and centrifuged within 30 min. Serum was obtained and kept frozen at - 20°C until assayed for androstenedione (A) and testosterone (T). Both hormones were quantified by radioimmunoassay after extraction and separation by celite column chromatography as described by Abraham et al. (12). [ 1,2,6,7-3H]testosterone (specific activity
*
*
of 94 Ci/mmol) and [1,2,6,7-3H]androstenedione (specific activity of 94 Cdmmol) were obtained from Amersham International (U.K.). Testosterone antiserum was kindly provided by the Human Reproduction Programme WHO (Switzerland), and the androstenedione antiserum was purchased from Radioassay Systems Lab (U.S.A.). All other reagents and solvents were of the highest purity available and purchased locally. The intra- and interassay coefficients of variation were 6.5 and lo%, respectively, for T, and 7.0 and 11%, respectively, for A. Data analysis was performed after logit-log transformation of the standard curve (13). Statistical analysis was performed by Student’s ‘‘t” test for unpaired observations. The interrelationship between A and T was evaluated by the linear regression method. p values less than 0.05 were considered to be significant. RESULTS
Androstenedioneand testosterone serum concentrations found in patients with pancreatic cancer and in the two control groups are presented in Table 1. Reference values and the calculated T1A ratio for each group are also included. In the pancreatic cancer group, androstenedione levels were significantly higher and testosterone levels were significantly lower when compared to patients with benign diseases or with other gastrointestinal malignancies. Consequently, the T/A ratio was also significantly lower in the pancreatic cancer group. There is a considerable overlap in the T and A levels between the groups; however, the highest A and lowest T values were found in patients with cancer of the pancreas (Figs. 1 and 2). No correlation was found between T and A levels (the “r” values are 0.28 for the pancreatic cancer group, 0.15 for benign
TABLE 1. Androstenedione and testosterone concentrations in patients with pancreatic carcinoma and in two control groups Androstenedione (pg/ml) Normal values (40-80 year old age group) Pancreatic cancer Benign (n = disease 39) (n = 37) Other gastrointestinal cancers (n = 36) All values are mean
f
4,597 f 2,034 1,834 2 2,113)=}
786 2 215 1,175 810Ia}
*
800 2 317
4,128
f
735
2
304
SD. N.S., not significant.
6.28 k 5 . 1 9 } }
N.S. 3,285
< 0.OOOl;’p
Testosterone1 androstenedione ratio
1.82 f 1.84
1,889
N.S.
u p = 0.01; ‘p = 0.003; ‘p = 0.000003;“p = 0.002; ‘p
Pancreas, Vol. 5 , No. 5 , 1990
Testosterone (pg/ml)
?
1,923
= 0.000005.
f
4.96
f
3.43
}
N.S.
-
517
PANCREATIC CANCER A N D ANDROGENS p=.003 oi
1
I
q
r
22
.DI
p = 0.000005 1 p=N.S. p=0.00001 -
a
a
:3000
.. ..... . .. ...... ..:h”.
i 4 0 0 1 I
H 2 000
gz 1000
”
I
PANCREATIC CANCER
0 0
18
p= N.S. I
I
0
. . . .
.*:.
*.: *-%+-
+%BENIGN OISEASES
.. . .
OTUERQICANCERS
FIG. 1. Serum androstenedioneconcentrationsin male patients with pancreatic carcinoma and in two control groups.
diseases, and 0.23 for the other gastrointestinal malignancies); however, the T/A ratio allowed the distinction of patients with cancer of the pancreas from other groups (Fig. 3). In two patients with stage I pancreatic cancer that was successfully resected, pre- and postoperative levels of A and T were obtained. The androgen profile changed markedly after surgery. Androstenedione concentrations dropped from 2,548 and 1,476 pg/ml to 1,180 and 143 pg/ml and testosterone values rose from 410 and 690 pg/ml to 5,750 and 5,959 pg/ml (Fig. 4). DISCUSSION
This study confirms previous findings of abnormalities in the androgen profile of male patients
FIG. 3. Serum testosterone/androstenedioneratio in male patients with pancreatic carcinoma and in two control groups.
with pancreatic carcinoma. We had previously described (7) that these patients had significantly lower T levels, and that its more active metabolite, 5-a-dihydrotestosterone (DHT), was also decreased, although not to the same extent, since the TlDHT ratio was found to be consistently below 5 (normal ratio = 10). This finding suggested that the tumor had a selective uptake of T as compared to DHT, or an increased conversion of T into DHT.
p = ,002
looo0lI
I
p= ,000003
r 90001 8 000
..
0
5000
0
I IOOO-
0.
*c
g
.. ..
p= N.S.
Androslonodlona 0----0
,
. ... .. .... .. . .’...
/4
5 000-
4 000-
3 000CASE Y 1 0 %
2 000-
IOOO-
.#*
I
BEFORE RESECTION
AFTER RESECTION
J
FIG. 2. Serum testosterone concentrations in male patients with pancreatic cancer and in two control groups.
FIG. 4. Preoperativeand postoperative serum androstenedione and testosterone concentrations in two patients with stage 1 pancreatic cancer that was successfully resected. Pancreas, Vol. 5 , No. 5 , 1990
518
C . FERNANDEZ-DEL CASTILLO ET AL,
This latter possibility is supported by the finding of increased levels of 5-a-reductase, the enzyme that converts T to DHT, in malignant pancreatic tissue
(14). In the present study, we confirm that low serum testosterone concentrations are found in patients with pancreatic cancer. Additionally, we demonstrate that androstenedione, the other major circulating C-19 steroid molecule with androgenic activity, is increased in these patients. In normal conditions, androstenedione is produced mainly by the adrenals (2-3 mg/day) and to a lesser extent by the testis (0.5 mg/day), where it is an intermediate metabolite in testosterone biosynthesis (15). We believe that the increased androstenedione found in pancreatic cancer patients is produced by the pancreatic tumor itself, and is not of adrenal or testicular origin. This possibility is supported by the fact that the enzyme that interconverts A and T is present in the canine pancreas (16). Furthermore, we have found that the isolated perfused canine pancreas readily converts T into A (10). If this conversion is present in the human and increased in neoplastic tissue, the low T and high A could easily be explained. The fact that concentrations of both hormones returned to normal after resection in two of our patients further supports this possibility and enables us to consider this abnormality as tumor marker in the diagnosis and followup of patients with pancreatic cancer resection. However, androstenedione itself appears to be of limited usefulness in the diagnosis of pancreatic cancer, since in most patients with this disease, concentrations of A fall within the physiological range. We consider that combining measurement of A with measurements of T, and thus obtaining the TIA ratio, will be of a greater clinical application since most of our males with cancer of the pancreas have abnormally low values. Further studies are justifiable to determine if the TIA ratio is useful in the detection and follow-up of patients with cancer of the pancreas. If pancreatic carcinoma is a sex steroid hormone dependent tumor, as has been suggested by the other authors (8,9), the finding of high androstenedione, probably due to increased testosterone transformation by the tumor, gives additional insight into the pathophysiology of this disease. Whether these changes are also present in female patients or rep-
Pancreas, Vol. 5 , No. 5, 1990
resent an indication of sexual dimorphism of the pancreas are possibilities that deserve further investigation. Acknowledgment:The authors wish to thank Dr. Guaddupe Garcia-Tsao for her kind review of the manuscript.
REFERENCES 1 . Connolly MM, Dawson PJ, Michelassi F, Moosa AR, Lowenstein F. Survival in 1001 patients with carcinoma of the pancreas. Ann Surg 1987;206:366-70. 2. Warshaw AL, Swanson RS.Pancreatic cancer in 1988. Possibilities and probabilities. Ann Surg 1988;208:541-53, 3. Levin DL, Connelly RR, Deveja SS. Demographic characteristics of cancer of the pancreas: mortality, incidence and survival. Cancer 1981;47:1456-68. 4. Czernichow P, Lerebours E, Hecktsweiler P, Colin R. Epid6miologie temporo-spatiale du cancer du pfincreas: etude de mortalit6 internationale et francaise. Gastroenterol Clin Biol 1985;9:767-75. 5 . Greenway B, Iqbal MJ, Johnson PJ, Williams R. Low serum testosterone concentration in patients with carcinoma of the pancreas. Br Med J 1983;286:93-5. 6. Shearer MG, Taggart D, Gray C, et al. Useful differentiation between pancreatic carcinoma and chronic pancreatitis by testosterone assay. Digestion 1984;30:106-7. 7. Robles-Diaz G, Diaz-Shchez V, Mtndez JP, Altamirano A, Wolpert E. Low serum testosterone/dihydrotestosteroneratio in patients with pancreatic carcinoma. Pancreas 1987;2: 3843. 8. Greenway B, Duke D, Pym B, Iqbal MJ, Johnson PJ, Williams R. The control of human pancreatic adenocarcinoma xenografts in nude mice by hormone therapy. Br J Surg 1982;69:595-7. 9. Lhoste EF, Roebuck BD, Stern JE, Longnecker DS. Effect of orchiectomy and testosterone on the early stages of azaserine induced pancreatic carcinogenesis in the rat. Pancreas 1987;2:38-43. 10. Fernfindez del Castillo C, Diaz-SBnchez V, Varela-Fascinetto G, Odor-Morales A, Altamirano A, Robles-Dtaz G. Testosterone biotransformation by the isolated perfused canine pancreas. Pancreas 1989;4615. 11. Beahrs OH, Henson DE, Hutter RVP, Myers MH, eds. Manual for staging of cancer, third edition. Philadelphia: J. B. Lippincott Co., 1988. 12. Abraham GE, Manhinos FS, Garza R. Radioimmunoassay of steroids. In: Abraham GE, ed. Handbook of radioimmunoassay. New York: Marcel Dekker Inc., 1977. 13. Rodbard D, Hutt MD. Statistical quality control and routine data processing for radioimmunoassay and immunoradiometric assays. Clin Chem 1974;20:125548. 14. Iqbal MJ, Greenway B, Wilkinson ML, Johnson PJ, Williams R. Sex steroid enzymes aromatase and 5 a reductase in the pancreas: a comparison of normal adult, foetal and malignant tissue. Clin Sci 1983;65:71-5. 15. Hudson B, Coghlan JP, Dulmanis A. Testicular function in man. In: Wolstenholme GEW, O’Connor M, eds. Endocrinology of the testis. London: J. & A. Churchill Ltd., 1967. 16. Mendoza-HemBndez G, Lbpez-Solache I, Rendon JL, DiazSfinchez V, Dtaz-Sagoya JC. 17~-hydroxysteroiddehydrogenase activity in canine pancreas. Biochem Biophys Res Commun 1988;152:376-82.
Pancreatic Cancer and Androgen Metabolism: High Androstenedione and Low Testosterone Serum Levels Carlos Ferniindez-del Castillo, Guillermo Robles-Diaz, "Vicente Diaz-SBnchez, and "Araceli Altamirano Pancreas Clinic and *Department of Reproductive Biology, Instituto Nacional de la Nutricih "Salvador Zubircin," Mexico City, Mexico
Summary: Serum androstenedione and testosterone levels were measured in 39 male patients with pancreatic cancer, and compared with the values obtained from 37 male patients with chronic pancreatitis or benign obstructive jaundice, and with those from 36 male patients with other gastrointestinal malignancies. Mean androstenedione values were significantly higher in the pancreatic cancer patients when compared to both control groups, and mean testosterone levels were significantly lower. The testosterone/androstenedione ratio was calculated and was also found to be significantly lower in the patients with pancreatic cancer. There was no difference in this ratio or in the androstenedione or testosterone levels when comparing both control groups. In two patients with stage I pancreatic cancer, serum androstenedione and testosterone levels were significantly altered, and returned to normal values after successful resection. These results confirm previous findings indicating that there is signifcant derangement in the androgen profile of patients with pancreatic cancer. Key Words: Pancreatic cancer-Androstenedione-TestosteroneAndrogens.
Cancer of the pancreas is a highly lethal disease whose frequency is increasing (1). Its high mortality is due to late diagnosis, and therefore many efforts are currently underway to detect the disease at earlier stages and to make a better differential diagnosis (2). Since pancreatic cancer affects more men than women (3,4), several clinical and experimental studies have focused on the possible role of androgens as indicators andor etiological factors in pancreatic carcinogenesis (5-9). Low serum testosterone levels have been found consistently in patients
with pancreatic carcinoma (5-7). Moreover, a low testosterone/dihydrotestosteroneratio has been described (7), suggesting an increased testosterone uptake by the tumor and increased transformation to the reduced metabolite, 5-a-dihydrotestosterone. In experimental models of pancreatic carcinogenesis, the tumor growth rate is stimulated by testosterone, whereas it is inhibited by antiandrogens or castration (8,9). These studies suggest that testosterone exerts atrophic effect on the pancreatic carcinoma. We have recently shown that the isolated per-
Address correspondence and reprint requests to Dr. G. Robles-Diaz at Department of Gastroenterology,Instituto Nacional de la Nutrici6n Sz, Vasco de Quiroga #is, nalpan 14000, ~ 6 ico D.F., Mexico.
hormones in male patients with cancer ~ and - in two Control groups (one with benign diseases of the pancreas or lower biliary tract, and the other
. ,
515
C . FERNANDEZ-DEL CASTILLO ET AL.
516
with gastrointestinal malignancies other than cancer of the pancreas).
PATIENTS AND METHODS Peripheral venous blood samples were obtained from 111 male patients who were divided into three groups. Group 1 comprised 39 patients with pancreatic carcinoma, aged 57 11 years (mean SD). The diagnosis of pancreatic cancer was histologically confirmed in 32 cases, and established by surgical exploration and compatible clinical course in the remaining 7 patients. According to AJCC staging (ll), seven patients were in stage I, two in stage 11, four in stage 111, and 26 in stage IV. Group I1 comprised 37 patients with benign pancreatobiliary disease, aged 54 & 14 years. Twentyeight had chronic pancreatitis, which was diagnosed on the basis of a compatible clinical picture and the presence of pancreatic calculi and/or abnormal exocrine function tests. The remaining nine patients had benign obstructive jaundice secondary to stones or bile duct stenosis. Group 111 comprised 36 patients with histologically proven gastrointestinal malignancies other than cancer of the pancreas, aged 58 ? 17 years. These carcinomas were distributed as follows: 15 gastric, 8 colorectal, 5 ampullary, 4 esophageal, 3 bile duct, and 1 hepatic. All blood samples were collected after an overnight fast and centrifuged within 30 min. Serum was obtained and kept frozen at - 20°C until assayed for androstenedione (A) and testosterone (T). Both hormones were quantified by radioimmunoassay after extraction and separation by celite column chromatography as described by Abraham et al. (12). [ 1,2,6,7-3H]testosterone (specific activity
*
*
of 94 Ci/mmol) and [1,2,6,7-3H]androstenedione (specific activity of 94 Cdmmol) were obtained from Amersham International (U.K.). Testosterone antiserum was kindly provided by the Human Reproduction Programme WHO (Switzerland), and the androstenedione antiserum was purchased from Radioassay Systems Lab (U.S.A.). All other reagents and solvents were of the highest purity available and purchased locally. The intra- and interassay coefficients of variation were 6.5 and lo%, respectively, for T, and 7.0 and 11%, respectively, for A. Data analysis was performed after logit-log transformation of the standard curve (13). Statistical analysis was performed by Student’s ‘‘t” test for unpaired observations. The interrelationship between A and T was evaluated by the linear regression method. p values less than 0.05 were considered to be significant. RESULTS
Androstenedioneand testosterone serum concentrations found in patients with pancreatic cancer and in the two control groups are presented in Table 1. Reference values and the calculated T1A ratio for each group are also included. In the pancreatic cancer group, androstenedione levels were significantly higher and testosterone levels were significantly lower when compared to patients with benign diseases or with other gastrointestinal malignancies. Consequently, the T/A ratio was also significantly lower in the pancreatic cancer group. There is a considerable overlap in the T and A levels between the groups; however, the highest A and lowest T values were found in patients with cancer of the pancreas (Figs. 1 and 2). No correlation was found between T and A levels (the “r” values are 0.28 for the pancreatic cancer group, 0.15 for benign
TABLE 1. Androstenedione and testosterone concentrations in patients with pancreatic carcinoma and in two control groups Androstenedione (pg/ml) Normal values (40-80 year old age group) Pancreatic cancer Benign (n = disease 39) (n = 37) Other gastrointestinal cancers (n = 36) All values are mean
f
4,597 f 2,034 1,834 2 2,113)=}
786 2 215 1,175 810Ia}
*
800 2 317
4,128
f
735
2
304
SD. N.S., not significant.
6.28 k 5 . 1 9 } }
N.S. 3,285
< 0.OOOl;’p
Testosterone1 androstenedione ratio
1.82 f 1.84
1,889
N.S.
u p = 0.01; ‘p = 0.003; ‘p = 0.000003;“p = 0.002; ‘p
Pancreas, Vol. 5 , No. 5 , 1990
Testosterone (pg/ml)
?
1,923
= 0.000005.
f
4.96
f
3.43
}
N.S.
-
517
PANCREATIC CANCER A N D ANDROGENS p=.003 oi
1
I
q
r
22
.DI
p = 0.000005 1 p=N.S. p=0.00001 -
a
a
:3000
.. ..... . .. ...... ..:h”.
i 4 0 0 1 I
H 2 000
gz 1000
”
I
PANCREATIC CANCER
0 0
18
p= N.S. I
I
0
. . . .
.*:.
*.: *-%+-
+%BENIGN OISEASES
.. . .
OTUERQICANCERS
FIG. 1. Serum androstenedioneconcentrationsin male patients with pancreatic carcinoma and in two control groups.
diseases, and 0.23 for the other gastrointestinal malignancies); however, the T/A ratio allowed the distinction of patients with cancer of the pancreas from other groups (Fig. 3). In two patients with stage I pancreatic cancer that was successfully resected, pre- and postoperative levels of A and T were obtained. The androgen profile changed markedly after surgery. Androstenedione concentrations dropped from 2,548 and 1,476 pg/ml to 1,180 and 143 pg/ml and testosterone values rose from 410 and 690 pg/ml to 5,750 and 5,959 pg/ml (Fig. 4). DISCUSSION
This study confirms previous findings of abnormalities in the androgen profile of male patients
FIG. 3. Serum testosterone/androstenedioneratio in male patients with pancreatic carcinoma and in two control groups.
with pancreatic carcinoma. We had previously described (7) that these patients had significantly lower T levels, and that its more active metabolite, 5-a-dihydrotestosterone (DHT), was also decreased, although not to the same extent, since the TlDHT ratio was found to be consistently below 5 (normal ratio = 10). This finding suggested that the tumor had a selective uptake of T as compared to DHT, or an increased conversion of T into DHT.
p = ,002
looo0lI
I
p= ,000003
r 90001 8 000
..
0
5000
0
I IOOO-
0.
*c
g
.. ..
p= N.S.
Androslonodlona 0----0
,
. ... .. .... .. . .’...
/4
5 000-
4 000-
3 000CASE Y 1 0 %
2 000-
IOOO-
.#*
I
BEFORE RESECTION
AFTER RESECTION
J
FIG. 2. Serum testosterone concentrations in male patients with pancreatic cancer and in two control groups.
FIG. 4. Preoperativeand postoperative serum androstenedione and testosterone concentrations in two patients with stage 1 pancreatic cancer that was successfully resected. Pancreas, Vol. 5 , No. 5 , 1990
518
C . FERNANDEZ-DEL CASTILLO ET AL,
This latter possibility is supported by the finding of increased levels of 5-a-reductase, the enzyme that converts T to DHT, in malignant pancreatic tissue
(14). In the present study, we confirm that low serum testosterone concentrations are found in patients with pancreatic cancer. Additionally, we demonstrate that androstenedione, the other major circulating C-19 steroid molecule with androgenic activity, is increased in these patients. In normal conditions, androstenedione is produced mainly by the adrenals (2-3 mg/day) and to a lesser extent by the testis (0.5 mg/day), where it is an intermediate metabolite in testosterone biosynthesis (15). We believe that the increased androstenedione found in pancreatic cancer patients is produced by the pancreatic tumor itself, and is not of adrenal or testicular origin. This possibility is supported by the fact that the enzyme that interconverts A and T is present in the canine pancreas (16). Furthermore, we have found that the isolated perfused canine pancreas readily converts T into A (10). If this conversion is present in the human and increased in neoplastic tissue, the low T and high A could easily be explained. The fact that concentrations of both hormones returned to normal after resection in two of our patients further supports this possibility and enables us to consider this abnormality as tumor marker in the diagnosis and followup of patients with pancreatic cancer resection. However, androstenedione itself appears to be of limited usefulness in the diagnosis of pancreatic cancer, since in most patients with this disease, concentrations of A fall within the physiological range. We consider that combining measurement of A with measurements of T, and thus obtaining the TIA ratio, will be of a greater clinical application since most of our males with cancer of the pancreas have abnormally low values. Further studies are justifiable to determine if the TIA ratio is useful in the detection and follow-up of patients with cancer of the pancreas. If pancreatic carcinoma is a sex steroid hormone dependent tumor, as has been suggested by the other authors (8,9), the finding of high androstenedione, probably due to increased testosterone transformation by the tumor, gives additional insight into the pathophysiology of this disease. Whether these changes are also present in female patients or rep-
Pancreas, Vol. 5 , No. 5, 1990
resent an indication of sexual dimorphism of the pancreas are possibilities that deserve further investigation. Acknowledgment:The authors wish to thank Dr. Guaddupe Garcia-Tsao for her kind review of the manuscript.
REFERENCES 1 . Connolly MM, Dawson PJ, Michelassi F, Moosa AR, Lowenstein F. Survival in 1001 patients with carcinoma of the pancreas. Ann Surg 1987;206:366-70. 2. Warshaw AL, Swanson RS.Pancreatic cancer in 1988. Possibilities and probabilities. Ann Surg 1988;208:541-53, 3. Levin DL, Connelly RR, Deveja SS. Demographic characteristics of cancer of the pancreas: mortality, incidence and survival. Cancer 1981;47:1456-68. 4. Czernichow P, Lerebours E, Hecktsweiler P, Colin R. Epid6miologie temporo-spatiale du cancer du pfincreas: etude de mortalit6 internationale et francaise. Gastroenterol Clin Biol 1985;9:767-75. 5 . Greenway B, Iqbal MJ, Johnson PJ, Williams R. Low serum testosterone concentration in patients with carcinoma of the pancreas. Br Med J 1983;286:93-5. 6. Shearer MG, Taggart D, Gray C, et al. Useful differentiation between pancreatic carcinoma and chronic pancreatitis by testosterone assay. Digestion 1984;30:106-7. 7. Robles-Diaz G, Diaz-Shchez V, Mtndez JP, Altamirano A, Wolpert E. Low serum testosterone/dihydrotestosteroneratio in patients with pancreatic carcinoma. Pancreas 1987;2: 3843. 8. Greenway B, Duke D, Pym B, Iqbal MJ, Johnson PJ, Williams R. The control of human pancreatic adenocarcinoma xenografts in nude mice by hormone therapy. Br J Surg 1982;69:595-7. 9. Lhoste EF, Roebuck BD, Stern JE, Longnecker DS. Effect of orchiectomy and testosterone on the early stages of azaserine induced pancreatic carcinogenesis in the rat. Pancreas 1987;2:38-43. 10. Fernfindez del Castillo C, Diaz-SBnchez V, Varela-Fascinetto G, Odor-Morales A, Altamirano A, Robles-Dtaz G. Testosterone biotransformation by the isolated perfused canine pancreas. Pancreas 1989;4615. 11. Beahrs OH, Henson DE, Hutter RVP, Myers MH, eds. Manual for staging of cancer, third edition. Philadelphia: J. B. Lippincott Co., 1988. 12. Abraham GE, Manhinos FS, Garza R. Radioimmunoassay of steroids. In: Abraham GE, ed. Handbook of radioimmunoassay. New York: Marcel Dekker Inc., 1977. 13. Rodbard D, Hutt MD. Statistical quality control and routine data processing for radioimmunoassay and immunoradiometric assays. Clin Chem 1974;20:125548. 14. Iqbal MJ, Greenway B, Wilkinson ML, Johnson PJ, Williams R. Sex steroid enzymes aromatase and 5 a reductase in the pancreas: a comparison of normal adult, foetal and malignant tissue. Clin Sci 1983;65:71-5. 15. Hudson B, Coghlan JP, Dulmanis A. Testicular function in man. In: Wolstenholme GEW, O’Connor M, eds. Endocrinology of the testis. London: J. & A. Churchill Ltd., 1967. 16. Mendoza-HemBndez G, Lbpez-Solache I, Rendon JL, DiazSfinchez V, Dtaz-Sagoya JC. 17~-hydroxysteroiddehydrogenase activity in canine pancreas. Biochem Biophys Res Commun 1988;152:376-82.
