ADULT UROLOGY
PROSTATE-SPECIFIC ANTIGEN IN FEMALE URINE:
A PROSPECTIVE STUDY INVOLVING 217 WOMEN
STEFAN SCHMIDT, MATTHIAS FRANKE, JAN LEHMANN, TILLMANN LOCH,
MICHAEL STÖCKLE, AND KLAUS WEICHERT-JACOBSEN
ABSTRACT
Objectives. Histomorphologic studies have provided evidence of prostate-specific antigen (PSA)-producing
tissue in the female urethra. Some urine samples from women in a small series were positive for PSA, but no
systematic investigation of this subject has been done to date.
Methods. In a prospective study, we analyzed whether PSA occurs in the urine of women and what factors
induce detectable PSA levels. The urine samples of 217 women were analyzed (Hybritech-Tandem E-PSA)
under standardized conditions. The impact of urine pH and volume was investigated, and the results were
correlated with clinical data (age, residual urine, urinary tract infection and prior sexual intercourse within
48 hours).
Results. A positive PSA level greater than the detection limit of 0.1 ng/mL was found in 11% of the analyzed
samples; their mean value was 0.29 ng/mL. pH correction did not result in a significant difference. The
voiding volume had no influence on the PSA level. Among the cases of detectable PSA, women younger than
50 years of age (n ⫽ 14) had a mean PSA of 0.34 ng/mL and those older than 50 years (n ⫽ 9) a mean of
0.23 ng/mL. One of 9 women with and 22 of 208 women without residual urine volume had a detectable PSA
level, as did 0 of 20 with and 23 of 197 women without urinary tract infection, and 3 of 7 with and 20 of 210
women without prior sexual intercourse within the previous 48 hours. None of the differences were
significant.
Conclusions. A urine PSA level was detected in 11% of all women studied, with PSA values apparently age
dependent. Any urine portion is suitable for analysis. No influence was determined for residual urine volume
or urinary tract infection. Sexual intercourse may cause detectable PSA values, but the data of this study did
not provide sufficient evidence for this hypothesis. UROLOGY 57: 717–720, 2001. © 2001, Elsevier Science
Inc.
F
emale periurethral glands exhibit analogies
with the male prostate. Secretion of prostatespecific antigen (PSA) is typical of this tissue.1 Histomorphologic studies on this prostate equivalent
have been known for some time under the term
“female prostate.”1– 4 In contrast to the intensive
research on serum PSA, urine PSA had not been
investigated more closely until the forensic study
by Graves et al.5 in 1985 in connection with rape
victims. Independently, Tremblay et al.6 addressed
the subject of measurement of urine PSA in pa-
tients with prostate cancer and benign prostatic
hyperplasia in 1987. Other investigators subsequently considered the role of urine PSA as a
screening parameter for prostate carcinoma recurrence.7–10 In these studies, female urine was used in
a very small number of cases for reference samples,
but did not provide any valid data on the detectability and significance of urine PSA in women, the
subject of the present standardized study.
From the Department of Urology, Christian-Albrechts-Universität zu Kiel, Kiel; and Department of Urology, Universitätskliniken des Saarlandes, Homburg, Germany
Reprint requests: Stefan Schmidt, M.D., Department of Urology, University of Kiel, Arnold-Heller-Strausse 7, 24105 Kiel,
Germany
Submitted: September 18, 2000, accepted (with revisions): November 22, 2000
Urine samples of 217 consecutive women were analyzed for
PSA values to determine what factors influence the PSA level.
All women admitted to our clinic within a period of 5 months
in 1998 were recruited. A minimum initial portion of morning
urine (volume in milliliters) was analyzed with the HybritechTandem E-PSA assay under the following standardized conditions: measurement in native urine (detection limit according
to the manufacturer of 0.1 ng/mL or greater) and after pH
© 2001, ELSEVIER SCIENCE INC.
ALL RIGHTS RESERVED
MATERIAL AND METHODS
0090-4295/01/$20.00
PII S0090-4295(00)01093-1 717
correction. Test sets for PSA determination in urine were not
commercially available. However, as Spitz et al.11 were able to
demonstrate, the estimation of PSA in urine can be performed
using a commercially available assay for serum specimens
without loss in accuracy. No modification was necessary.11
LABORATORY TECHNIQUES
The samples were stored for a maximum period of 2 weeks
at ⫺20°C before measurement. All urine samples underwent
agitation for mixing purposes, after which the pH was measured and 1 mL removed for PSA analysis. The remaining
urine was titrated to pH 7.5, after which another sample was
taken.
The correlation of the urine PSA values with pH, urine volume, age, residual urine volume, urinary tract infection and
sexual intercourse within 48 hours before sampling was then
determined in statistical terms. The distribution type of the
measurements was assessed by the Kolmogorov-Smirnov test.
To determine the significance of the correlation among urinary PSA, pH, and volume, the Wilcoxon-Mann-Whitney test
was applied. Furthermore, the Fisher exact test was performed
for statistical differences between the clinical subgroups.
RESULTS
PSA (greater than 0.1 ng/mL) could be detected
in the urine of 23 (11%) of 217 women; the mean
was 0.29 ng/mL (median 0.19, range 0.12 to 1.06)
with an average urine pH of 6.2 (range 5.0 to 8.5).
The mean PSA level in all native urine samples was
0.047 ng/mL (SD 0.08). Correcting the pH to pH
7.5 resulted in a change in urine PSA levels to
0.050 ng/mL (SD 0.11; difference not significant).
The study parameters were thus clearly not pH dependent. The influence of volume of the spontaneous urine on urine PSA was also not significant. In
volumes smaller than 50 mL, the average PSA level
was 0.056 ng/mL (SD 0.12) (n ⫽ 94); at volumes of
50 mL or greater, the average PSA level was 0.058
ng/mL (SD 0.12) (n ⫽ 123). Nondetectable PSA
values were regarded as 0.00 ng/mL. In terms of
volume, 13 (14%) of 94 samples (mean PSA 0.222
ng/mL; volume less than 50 mL) and 10 (8%) of
123 samples (mean PSA 0.389 ng/mL; volume 50
mL or greater) were PSA positive. The average PSA
concentration in the patient group with larger volumes exceeded the PSA concentration in the patient group with smaller volumes by a factor of 1.7,
thus ruling out a dilution effect. The correlation
between age and urine PSA (age range 18 to 74
years, mean age 43, median age 42) revealed a statistically insignificant tendency toward a reduction
in PSA level with increasing age. Women younger
than 50 years of age (n ⫽ 14) had a mean PSA level
of 0.34 ng/mL (median 0.19); those older than 50
(n ⫽ 9) had a mean PSA level of 0.23 ng/mL (median 0.15). A PSA-positive result was obtained in 1
of 9 with and 22 of 208 without a residual urine
volume, 0 of 20 with and 23 of 197 without urinary
tract infection, and 3 of 7 women with and 20 of
718
210 women without prior sexual intercourse, all
without statistical relevance.
COMMENT
In this study, PSA in female urine, along with the
question of its dependence on various factors, was
investigated systematically and prospectively in a
large group of women for the first time. PSA was
found in the urine of every ninth woman. Histologic analyses of the female urethra by several
other investigators have provided evidence for the
potential origins of such PSA-positive cases. Pollen
and Dreilinger2 detected the presence of PSA-positive cells in 7 of 10 female periurethral glands by
immunohistochemical methods in 1984. Wernert
et al.1 found PSA-positive histologic features in two
thirds of a collective of 33 women. Since the embryologic origin of the periurethral glands corresponds to that of the male prostate, a histomorphologic similarity is also probable.4 PSA expression
was also detected in the urachal structures in autopsies of both men and women.3 A comparison
with the few studies on PSA measurement in the
urine of small collectives of women revealed varying results: Takayama et al.9 (n ⫽ 12) and Tremblay et al.6 found no PSA in female urine. Both Spitz
et al.11 (n ⫽ 8) and Graves et al.5 (n ⫽ 15) detected
no levels above the detection limit of 0.3 ng/mL
and 3 ng/mL, respectively. De Vere White et al.,7 in
contrast, detected an average level of 0.2 ng/mL
PSA in 10 of 15 women, in which sexual intercourse was not taken into account. Breul et al.10
found PSA levels exceeding 0.5 ng/mL in one third
of his collective (n ⫽ 34), averaging 3.72 ng/mL;
the levels in women after sexual intercourse were
clearly higher (31 ng/mL). In a recent study, Breul
et al.12 detected an average level of 1.73 ng/mL in
20 women. Applying a sensitive test to our much
larger collective of 217 women, we detected PSA in
the urine of 11%. In these women, the average level
(0.29 ng/mL) was considerably greater than the
detection limit. This small percentage of PSA-positive urine samples was probably because we evaluated a nonselected collective.
As with Spitz et al.,11 and in contrast to the other
investigators,7,9,10 we were also unable to determine any influence of pH shifts. pH correction appeared to be of no value for PSA measurement in
urine.
In their studies of male patients, Takayama et
al.,9 Iwakiri et al.,8 and Tremblay et al.6 concluded
that PSA determination in first-stream urine samples leads to higher results. Likewise Spitz et al.11
also reported this tendency in their measurements
of fractionated spontaneous urine in men. In their
study, the first 40-mL portion was characterized by
PSA concentrations about two to four times higher
UROLOGY 57 (4), 2001
than those in subsequent urine fractions, except
for the last portion, which showed a renewed increase. This evidence supports the hypothesis that
PSA is produced by the periurethral glands in the
distal urogenital tract, where it is flushed out by the
urine. In this manner, a dilution effect, as Spitz et
al. have found, can be expected. The increase of
PSA in the last portion is interpreted as being due
to the muscle contraction of the bladder neck and
urethra at the end of voiding.11 According to the
hypothesis that PSA is produced by the periurethral glands, it was confirmed in studies of renal
fistular urine samples10,11 that the urine PSA concentration is not dependent on serum PSA. Renal
dysfunction also has no influence on the serum
value.13,14 The measurable serum level in women
varies between 0.0 and 0.2 ng/mL (the “female
range”).15,16 Conversely, PSA secretion of the periurethral glands (potential “leak back”) also has no
influence on the serum concentrations.17 The
above-mentioned dilution effect was absent in our
findings in female urine. Without exception, our
analyses were carried out in first-stream urine, in
which no clinically relevant dilution effect in urine
samples up to a maximum micturition volume of
100 mL could be detected. The difference seems to
depend on the known diverse morphologic findings in the male and female urethra. The emission
of PSA by the periurethral glands during voiding
appears to be different in men and women. Moreover, the average PSA level in the urine of men
(median 10, 20, and 30 ng/mL,18 median 300 ng/
mL, range 6 to 442 ng/mL11) was found to be 30 to
1000 times higher than in the urine of women investigated in our department (0.29 ng/mL). Accordingly, in healthy men, the urinary PSA rates
can be reduced to the female values by antiandrogenic therapy (untreated men median 10 to 30
ng/mL versus treated men 0.09 to 1.1 ng/mL, mean
0.251 ng/mL with cyproterone acetate).18 Likewise, androgen administration in women leads to
increased urinary PSA levels (baseline level mean
0.017 ng/mL versus 18.130 ng/mL after testosterone treatment), so that the urinary PSA measurements may be useful as indicators of hyperandrogenism in women.19
Data on the age-dependent distribution of PSA
concentration in women vary. On one hand, the
female urethral glands remain rudimentary during
the aging process (ie, they do not reach the histomorphologic change observed in men),1 correlating with the tendency toward lower urine PSA levels in older women compared with younger
women. In accordance with our finding, Yu et al.20
found in an immunoreactive study of PSA tissue
expression in breast tumors that PSA positivity declines with age, in conformity with a negative association between tissue PSA and age. The same
UROLOGY 57 (4), 2001
observation was made by Melegos et al.21 in female
serum of hirsute women. On the other hand, extraprostatic PSA production would appear to be under
hormonal, androgen-dependent control.12,18 –21 A
corresponding study by Yu and Diamandis15 on female sera (n ⫽ 1064) revealed just the opposite tendency (ie, toward a PSA distribution pattern with
higher levels in old age). Yu and Diamandis15 found
that the menopausal status of these older women was
responsible for the production of PSA. PSA production is known to be upregulated by androgen and
suppressed by estrogen; after menopause, the ratio of
estrogen to androgen declines, a situation that may
favor PSA production.
The periurethral glands may be affected by urinary tract infections.22 Our study did not reveal a
corresponding correlation in the form of a PSA
level shift in urinary tract infections analogous to
the frequent rise in serum PSA in prostatitis. Similarly, no evidence was found of a relation to residual urine status.
Graves et al.5 investigated the suitability of urine
PSA in women relevant to the forensic aspects of
sexual crimes. In addition to tissue-produced PSA,
sexual intercourse is a possible cause of PSA in
urine, because sperm contains high concentrations
of PSA.5 Adult male volunteers were urine positive
in this study and female volunteers were negative.
The test used was not highly sensitive, having a less
sensitive detection limit of 3.0 ng/mL.5 In contrast
to the PSA levels measured in the urine after sexual
intercourse in the studies of other investigators,
some of them high,7,10 Graves et al., as well as ourselves, were unable to demonstrate an influence of
sexual intercourse on PSA concentration or frequency in the urine of women. Of 7 women who
had had sexual intercourse within 48 hours before
sampling, only 3 had measurable positive PSA levels.
The question as to whether urine PSA might be
an effective tumor marker has been investigated in
many studies of men.6 –10 Although De Vere White
et al.7 found raised PSA levels more frequently in
the midstream urine of prostatectomy patients
than in serum and Tremblay et al.6 reported higher
urine values than in healthy persons, the role of
urine PSA as a screening parameter for tumors remains unclarified. PSA in the urine after prostatectomy could be caused by either residual prostate
tissue or a local recurrence, or it could reflect the
production of the urethral glands.7–10 In women,
there are case reports of the rare adenocarcinoma
of the female urethra originating in Skene’s
glands,23 which is PSA positive according to Wernert et al.1 and others.24,25 Additionally, a case report
of Skene’s gland adenocarcinoma showed an increased serum level of PSA preoperatively, which
promptly decreased after surgical excision of the
719
lesion.26 Because of the rarity of this tumor, the
presence of positive PSA in female urine on the
basis of adenocarcinoma of the urethra would be
very hard to prove.
CONCLUSIONS
In a unselected collective of urologic patients, 1
woman in 9 was positive for PSA in the urine. The
finding seemed to be dependent on age and was
without a dilution effect in small volumes. Influences arising from the factors pH, sexual intercourse, residual urine, and urinary tract infection
were not evident.
REFERENCES
1. Wernert N, Albrecht M, Sesterhenn I, et al: The “female” prostate: location, morphology, immunohistochemical
characteristics and significance. Eur Urol 22: 64 – 69, 1992.
2. Pollen JJ, and Dreilinger A: Immunohistochemical
identification of prostatic acid phosphatase and prostatic specific antigen in female urethral glands. Urology 23: 303–304,
1984.
3. Golz R, and Schubert GE: Prostatic specific antigen:
immunoreactivity in urachal remnants. J Urol 141:
1480 –1482, 1989.
4. Tepper SL, Jagidar J, Heath D, et al: Homology between
the female paraurethral (Skene’s) glands and the prostate.
Arch Pathol Lab Med 108: 423– 425, 1984.
5. Graves H, Sensabaugh GF, Crim D, et al: Postcoital
detection of a male specific semen protein. N Engl J Med 312:
338 –343, 1985.
6. Tremblay J, Frenette G, Tremblay RR, et al: Excretion of
three major prostatic proteins in the urine of normal men and
patients with benign hypertrophy or prostate cancer. Prostate
10: 235–243, 1987.
7. De Vere White RW, Meyers JF, Soares SE, et al: Urinary
prostate specific antigen levels: role in monitoring the response of prostate cancer to therapy. J Urol 147: 947–951,
1992.
8. Iwakiri J, Grandbois K, Wehner N, et al: An analysis of
urinary prostate specific antigen before and after radical prostatectomy: evidence for secretion of prostatic specific antigen
by the periurethral glands. J Urol 149: 783–786, 1993.
9. Takayama TK, Vessela RL, Brawer M, et al: Urinary
prostate specific antigen levels after radical prostatectomy.
J Urol 151: 82– 87, 1994.
10. Breul J, Pickl U, and Hartung R: Prostate specific antigen in urine. Eur Urol 26: 18 –21, 1994.
11. Spitz J, Bickert T, and Köllermann M: Technische und
720
physiologische Aspekte der Bestimmung des PSA-Spiegels im
Urin. Klin Lab 41: 449 – 455, 1995.
12. Breul J, Pickl U, and Schaff J: Extraprostatic production
of prostate specific antigen is under hormonal control. J Urol
157: 212–213, 1997.
13. Morton JJ, Howe SF, Lowell JA, et al: Influence of endstage renal disease and renal transplantation on serum prostate specific antigen. Br J Urol 75: 498 –501, 1995.
14. Kabalin JN, and Hornberger JC: Prostate specific antigen is not excreted by human kidney or eliminated by routine
hemodialysis. Urology 37: 308 –310, 1991.
15. Yu H, and Diamandis EP: Measurement of serum prostate specific antigen levels in women and in prostatectomized
men with an ultrasensitive immunoassay technique. J Urol
153: 1004 –1008, 1995.
16. Oesterling JE, Chan DW, Epstein JI, et al: Prostate specific antigen in the preoperative and postoperative evaluation
of localized prostate cancer treated with radical prostatectomy. J Urol 139: 766 –768, 1988.
17. Oesterling JE, Martin SK, Bergstrahl EJ, et al: The periurethral glands do not have a clinically significant effect on the
serum PSA concentration (abstract). J Urol 153(Pt 2): 518A,
1995.
18. Obiezu CV, Giltay EJ, Magklara A, et al: Dramatic suppression of plasma and urinary prostate specific antigen and
human glandular kallikrein by antiandrogens in male-to-female transsexuals. J Urol 163: 802– 805, 2000.
19. Obiezu CV, Giltay EJ, Magklara A, et al: Serum and
urinary prostate-specific antigen and urinary human glandular kallikrein concentrations are significantly increased after
testosterone administration in female-to-male transsexuals.
Clin Chem 46: 859 – 862, 2000.
20. Yu H, Diamandis EP, and Sutherland DJ: Immunoreactive prostate-specific antigen levels in female and male breast
tumors and its association with steroid hormone receptors and
patient age. Clin Biochem 27: 75–79, 1994.
21. Melegos DN, Yu H, Ashok M, et al: Prostate-specific
antigen in female serum, a potential new marker of androgen
excess. J Clin Endocrinol Metab 82: 777–780, 1997.
22. Huffman JW: Clinical significance of the paraurethral
ducts and glands. Arch Surg 62: 615– 626, 1951.
23. Knoblich R: Primary adenocarcinoma of the female
urethra. Am J Obstet Gynecol 80: 353–364, 1950.
24. Murphy DP, Pantuck AJ, Amenta PS, et al: Female urethral adenocarcinoma: immunohistochemical evidence of
more than 1 tissue of origin. J Urol 161: 1881–1884, 1999.
25. Sloboda J, Zaviacic M, Jakubovsky J, et al: Metastasizing
adenocarcinoma of the female prostate (Skene’s paraurethral
glands): histological and immunohistochemical prostate
markers studies and ultrastructural observation. Pathol Res
Pract 194: 129 –136, 1998.
26. Dodson MK, Cliby WA, Keeney GL, et al: Skene’s
glands adenocarcinoma with increased serum level of prostate-specific antigen. Gynecol Oncol 55: 304 –307, 1994.
UROLOGY 57 (4), 2001