0022-534 7/92/14 73-0810$03.

00 /0
THE JOURNAL OF UROLOGY Vol. 147, 810-814, March 1992
Copyright© 1992 by AMERICAN UROLOGICAL ASSOCIATION, INC, Printed in U.S.A.

Diagnosis

EFFECTS OF RECTAL EXAMINATION, PROSTATIC MASSAGE,

ULTRASONOGRAPHY AND NEEDLE BIOPSY ON SERUM PROSTATE
SPECIFIC ANTIGEN LEVELS
JERRY J. J. YUAN, DOUGLAS E. COPLEN, JOHN A. PETROS, ROBERT S. FIGENSHAU,
TIMOTHY L. RATLIFF, DEBORAH S. SMITH AND WlLLlAM J. CATALONA*
From the Division of Urologic Surgery and Division of Biostatistics, Washington University School of Medicine, St. Louis, Missouri

ABSTRACT

Measurement of serum prostate specific antigen (PSA) is commonly used to evaluate the prostate
gland in a variety of clinical settings. We examined the effects of pro static manipulations, including
digital rectal examination, prostate massage, transrectal ultrasonography and transrectal needle
biopsy, on serum PSA levels in 199 men. We detected no clinically significant difference between
serum PSA levels obtained immediately before and at 5 or 90 minutes after rectal examination in
43 men. We observed falsely increased PSA levels (to greater than 4 ng.fml., Tandem-R) in 1 of 17
men (6%) following prostatic massage and in 3 of 27 men (11%) following ultrasonography.
Transrectal needle biopsy caused an immediate increase in serum PSA in 92 of 100 men. In 29 of
these 92 men (32%) when followed weekly serum PSA levels did not return to baseline as expected
according to the published serum PSA half-life of 2 to 3 days. Biopsies taking 3 or fewer cores (7
patients) resulted in a smaller increase in serum PSA (mean 1.63 ± 1.12 times the baseline level
versus 6.24 ± 1.10 times baseline, p <0.03) and a proportionally shorter duration of PSA elevation
(mean 1.43 ± 0.48 weeks versus 2.13 ± 0.14 weeks, p = 0.20) than those taking 4 or more cores (93
patients). Prostate size and the presence of cancer had no influence on the duration of PSA elevation
following biopsy.
We conclude that digital rectal examination, prostatic massage and ultrasonography have minimal
effects on serum PSA levels in most patients. However, prostatic needle biopsy usually causes
marked elevations of serum PSA levels with a persistent PSA leak into the blood stream lasting
longer than expected from the serum half-life of PSA in approximately 25% of the patients.
KEY WORDS: prostate; massage; ultrasonic diagnosis; biopsy, needle; antigens, differentiation

Prostate specific antigen (PSA) is a serine protease of the vealed that despite a direct relationship between serum PSA
kallikrein family with a molecular weight of 35,000 daltons level and pathological tumor stage, substantial overlap of PSA
which serves to liquefy the seminal coagulum. PSA is produced levels existed among patients with different tumor stages, thus
exclusively by the prostatic ductal and acinar epithelium, and rendering serum PSA less useful as a staging tool. 4 Similar
it is composed of a single polypeptide chain of 240 amino results were reported by others. 8
acids. 1- 3 Elevated serum PSA levels have been associated with We are conducting an ongoing study of PSA as a first-line
benign prostatic hyperplasia (BPH), prostatitis and prostate screening test for prostate cancer and we have observed recently
cancer. The role of serum PSA as a tumor marker has yet to that in surgically staged cases a baseline serum PSA level of
be fully defined. Serum PSA is useful in monitoring the re- greater than 10 ng./ml. (Tandem-R) was a surprisingly good
sponse to treatment in patients with prostate cancer. 4- 9 The predictor of extracapsular tumor extension when there had
usefulness of serum PSA in the preoperative staging of cases been no recent prostate manipulation. 11 This prompted a ret-
of clinically localized prostate cancer is controversial. Stamey rospective review of 30 patients who had been evaluated before
et al reported that a preoperative serum PSA level of greater the screening study. These patients had pathologically organ-
than 50 ng./ml. (Yang Pros-Check polyclonal radioimmuno- confined prostate cancer but serum PSA levels greater than 10
assay) is a reliable predictor of extracapsular tumor extension. 5 ng./ml. preoperatively. We found that 75% of these patients
In contrast, Partin et al suggested that serum PSA failed to had had either prostate biopsies or transurethral resections
reflect accurately tumor burden and final pathological stage in performed within 30 days before the PSA levels were drawn, or
individual patients. 10 Our previous experience using the Tan- they had rectal examinations performed on the same day that
dem-R (Hybritech) monoclonal immunoradiometric assay re- the PSA levels were measured, which suggests that prior pros-
tatic manipulations may have spuriously increased the serum
Accepted for publication October 18, 1991. PSA levels in some of the patients with organ-confined cancer.
Editor's note. This paper was awarded a prize in the 1990 essay To examine this issue further, we conducted the present study
contest sponsored by the American Urological Association.
Supported in part by Hybritech, San Diego, California. on the effects of digital rectal examination, prostate massage,
* Requests for reprints: 4960 Audubon Ave., St. Louis; Missouri transrectal ultrasonography and transrectal needle biopsy of
63110. the prostate on serum PSA levels.
810
 EFFECTS OF PROSTATl:C MANIPULATIONS ON SERUM: PROSTATE SPECIF'IC ANTIGEN LEVELS 811
MATERIALS AND METHODS in 36 men. We selected the highest pre-uitrasonography serum
PSA level as the baseline. Serum PSA levels were obtained 5
We used the Tandem-R monoclonal i:mmunoradiometric
minutes after ultrasonography and we took additional meas-
PSA assay (Hybritech). The normal range recom~ende·d· by urements in 12 of these 36 men 24 hours later to assess a
the manufacturer is Oto 4 ng./ml. We performed routme d1g1tal
possible delayed increase. We established baseline seru~ PSA
rectal examination to define prostate size and asymmetry, and
levels in 100 men according to the same procedure used m the
to evaluate induration or fixation. One of us (J. J. J. Y.)
ultrasonography group. Serum PSA levels were obtained 5
performed prostate massage for 30 seconds. as previously de-
minutes after needle biopsies and then at weekly intervals until
scribed. 12 We performed ultrasonography w1~h 7 MHz. tra~s-
the PSA levels had returned to baseline. We applied the same
ducer scanning in the transverse and the sag1ttal planes, usmg
criterion to subjects with normal baseline PSA levels (4 ng./
either the Bruel and Kjaer model 1846 or the Teknar Proscan.
ml. or less), even if the returned serum PSA was greater than
We used an automatic biopsy gun fitted with an 18 gauge needle
for transrectal biopsy under sonographic guidance. We studied 4 ng./ml. For the first 23 men we per.formed.~u.ltiple determi-
nations at 48 to 72-hour intervals durmg the m1tial 7 to 10 days
199 ambulatory men who were evaluated at our outpatient
following biopsies. These multiple determinations added ~o
office. We excluded those with known or subsequently proved
further information to this study and were subsequently drn-
metastatic prostate cancer and suspected acute bacterial pros-
continued.
tatitis and those who had undergone prostatic needle biopsy
We used the 2-sample t test and paired t test for statistical
or tra~surethral surgery within 6 months, or rectal examination
analysis. All values were given as mean ± 1 standard error.
within 7 days.
PSA is known to fluctuate spontaneously from day to day, a RESULTS
phenomenon that we called biological variation in this study.
We calculated the biological variation in 2 ways. We selected Rectal examination had no clinically significant effect on
100 biopsied cases for this analysis; for each patient we per- serum PSA levels either immediately or 90 minutes after the
formed at least 2 pre-biopsy PSA levels (PSAl and PSA2) on examination. Three men in the 5-minute group had PSA ele-
different days approximately 2 weeks apart. We separ~ted them vations from baselines of 0.8 to 1.6 ng./ml., 1.5 to 2.2 ng./ml.
into 3 groups: normal PSA (O to 4 ng./ml.), low titer P?A and 2.9 to 3.6 ng./ml., respectively, while 1 man in the 90-
elevation (4.01 to 10 ng./ml.) and high titer PSA elevat10n minute group had an increase from 4.9 to 7.9 ng./ml. Of. 20
(greater than 10 ng./ml.). In the first calculation, called the men 3 (15%) had elevations in serum PSA levels followmg
difference :score, for each PSA pair (PSAl and PSA2) we prostatic massage from 3.6 to 4.8 ng./ml., l.~ to 3.3 ng./ml. ~nd
calculated the individual difference score (PSA2 minus PSAl) 1.5 to 2.2 ng./ml., respectively. Of 17 men with normal baselme
and then the mean group difference score and the correspond- PSA levels 1 (6%) had an abnormal level following prostatic
massage.
ing 99% confidence interval for each PSA group. We used the
formula: confidence interval = mean group difference score ± Of 36 men 4 (11 %) had elevations of PSA levels following
t 0.01 (standard deviation/-Jsample size -1 ). Confidence inter- ultrasonography. The first 3 men all coincidentally had baseline
vals were based upon the t distribution due to small sample PSA levels of 3.8 ng./ml. with subsequent increases to 6, 13.9
sizes. We defined the allowable biological variation as the upper and 14.1 ng./ml., respectively. In 2 men 24-hour measurements
limit of the appropriate confidence interval. In t~e ~e?ond were available: the PSA level in 1 patient had normalized and
calculation of biological variation we calculated the md1v1dual the level in the other patient remained elevated at 4.7 ng./ml.
coefficient of variation for each PSA pair and then the mean The fourth patient had an increase from 9.1 to 11.5 ng./ml. No
coefficient of variation for each PSA group. Two times. the significant change occurred in the remaining 32 men following
mean coefficient of variation times baseline was then defmed ultrasonography. None of the 12 men from whom additional
as allowable biological variation for each group. The biologi.cal 24-hour post-ultrasonography measurements were obtained
variations based on the difference score and the correspondmg had a delayed increase in PSA level. Overall 3 of 27 men (11.%)
with normal baseline PSA levels had abnormal levels followmg
mean group coefficients of variation. are sh~wi:i in t.able 1. The
2 methods for calculating the biological vanat10n yielded com- ultrasonography.
Table 2 summarizes the clinical information on the biopsied
parable results. . .
We defined an elevation of serum PSA followmg pros ta tic patients. In 92 of 100 men (92%) PSA levels were elevated
manipulation as a PSA increase exceeding. the allowable bio- immediately following transrectal biopsy. The immediate in-
logical variation; we defined return to baselme as. the ~eturn ~f crease of serum PSA following needle biopsies varied from none
an elevated serum PSA to within the allowable b10log1cal ".an- to 52-fold, with a mean of 5.91 times the baseline leveL Biopsies
ation over baseline. In this report we included only elevat10ns taking 3 or fewer cores caused a smaller initial increase in the
and returns to baseline fulfilling difference score and the coef- serum PSA level than those taking 4 cores or more (1.63 ± 1.12
ficient of variation criteria. times the baseline level versus 6.24 ± 1.10 times baseline, p
Serum PSA levels were obtained in 43 men immediately <0.03). The immediate post-biopsy serum PSA level repre-
before rectal examination. We measured post-rectal examina- sented the highest overall post-biopsy level in the great majority
tion levels 5 minutes after examination in 23 men and 90 of men studied. Following needle biopsies we performed serial
minutes after rectal examination in the remaining 20. Serum PSA measurements at 48 to 72-hour intervals in 23 men, of
PSA levels were obtained in 20 men immediately before pros- whom 1 had an additional increase surpassing the immediate
tatic massage. We measured post-massage levels 5 m~nutes post-biopsy value (22.3 ng./ml. at 48 hours versus 14 ng./ml. 5
after massage in 10 men and 90 minutes after massage m. the minutes after biopsy). Of 77 men who were tested on a weekly
remaining 10. Two ambulatory serum PSA levels were obtamed schedule 2 (3%) had delayed increases 1 week after biopsy. We

TABLE 2. Clinical data on biopsied patients

TABLE 1. Biological variation of serum PSA levels calculated by the Baseline PSA level (ng./ml.):
difference score and coefficient of variation determined in 100 study 0 to 4 27
patients 4.01 to 10 58
Greater than 10 15
Baseline PSA Difference Score Coefficient of Mean No. biopsies (range)
No. 4.8 (2-8)
(ng./ml.) (ng./ml.) Variation(%)
Mean age (range) 64 (51-78)
0-4 27 0.5 6.8 Biopsy result:
4.01-10 58 1.3 12.1 Ca 33
Greater than 10 15 5.8 9.7 Benign 67
-----
812 YUAN AND ASSOCIATES

separately studied 10 men with daily PSA levels for 2 to 3 days localized prostate cancer is controversial. Although it is gener-
following transrectal biopsy and found no additional increase ally agreed that a direct correlation exists between increasing
in serum PSA level surpassing the immediate post-biopsy value. serum PSA levels and higher pathological tumor stages, many
Of the 92 men with PSA elevations immediately following investigators have reported that the PSA levels for different
biopsy 29 (32%) had persistent serum PSA elevations 2 weeks pathological stages can overlap to a significant degree. Because
after biopsy (tables 3 and 4). Biopsies taking 3 or fewer cores investigators were unable to isolate an exact relationship be-
resulted in proportionally shorter duration of PSA elevations tween a given PSA level and a given tumor stage, many believe
than those taking 4 or more cores (1.43 ± 0.48 weeks versus that serum PSA was not accurate enough to be clinically useful
2.13 ± 0.14 weeks, p = 0.20). Prostate size (less than 30 cc, 30 in preoperative staging. 4 • 3- 10 However, we have observed that
to 60 cc and greater than 60 cc) had no apparent influence on in the absence of recent prostate manipulation a baseline serum
the duration of PSA elevation (2.15 ± 0.32 weeks, 2.26 ± 0.21 PSA level of greater than 10 ng./ml. predicts extracapsular
weeks and 2.01 ± 0.17 weeks, respectively, p >0.20); neither tumor extension or positive surgical margins in two-thirds of
did the presence of carcinoma (2.00 ± 0.27 weeks with carci- prostate cancer cases. These findings contradict most previ-
noma versus 2.42 ± 0.35 weeks with benign biopsy, p = 0.35). ously published reports, 3- 10 including our own. 4 The discrepancy
The_ figure illustrates deJay_ed PSAreturn to baseline following may be due to the IH'_e&nce of spuriously high_ PSa _levels in
biopsy (patient J. G., table 4). The serum half-life of PSA has blood samples that were drawn too soon after prostatic biopsy
been reported to be 2.2 days and 3.15 days following total in previous studies. To provide an overview of previous findings,
prostatectomy. 9 • 12 We used the latter half-life in this report. table 5 summarizes the available studies on the effects of
various prostatic manipulations on serum PSA levels.
DISCUSSION Serum PSA levels have been reported to have no diurnal
Serum PSA levels are measured to monitor prostate cancer variation. 13• 14 Schifman et al reported a 6.2% variation ex-
after therapy. Currently, many urologists believe that this is pressed as the coefficient of variation (standard deviation/
the extent of the usefulness of this serum marker. Thus, the mean) in 10 cases sampled twice daily for 3 days. 13 El-Shirbiny
use of serum PSA level in the preoperative staging of clinically et al noted a 9.9% variation in 16 men with hourly serum PSA
measurements for 24 hours. 14 These studies suggest that the
TABLE 3. PSA return to baseline following biopsy
biological variation of PSA measurements is approximately
10%, which parallels our calculation of 6.8 to 12.1 %, depending
No, Men
upon the baseline PSA level. Our results are in agreement with
No elevation 8 those of Brawer et al who found that digital rectal examination
Time to baseline (wks.): has minimal effects on serum PSA levels. 15 In a study involving
1 35
2 28 24 men they detected no significant difference among serum
3 14 PSA levels obtained 40 and 10 minutes before and those ob-
4 or more 15 tained 5 and 30 minutes after rectal examination. Similarly,
Total 100
using a sample of 43 men we detected no clinically significant

TABLE 4. Followup summary in 29 men with persistently elevated serum PSA following needle biopsy
Wks. of PSA (ng./ml.)
Pt. No. Biopsies Biopsy Results
Elevation* Baseline Post-Biopsy 1 Wk. 2Wks. 3 Wks. 4 Wks. 5 Wks. 6 Wks.

Normal baseline PSA

CM 5 Prostate Ca 4 3.0 18.9 14.3 6.1 8.6 4.9 3.5t
OA 5 Benign 4+ 3.0 56.4 6.1 4.8 4.1 4.0
JT 4 Benign 6+ 3.1 26.8 7.0 5.9 5.7 5.6 5.4 4.4
PHe 3 Prostate Ca 3 3.2 4.8 4.7 4.1 3.4t
KN 5 Benign 3 3.3 32 12 5.7 3.4t
CV 4 Benign 3 3.5 46.4 7.0 5.6 4.1
RW 6 Benign 5+ 3.6 164 13.6 11.0 7.3 7.3 5.5
GLe 4 Benign 4+ 3.8 46.7 7.0 6.9 5.9 5.6
PHi 6 Prostate Ca 3+ 3.9 30.5 8.2 6.2 7.8
RF 6 Prostate Ca 5 3.9 13.1 8.5 6.2 7.0 4.7 4.2t
ER 4 Benign 4 3.9 21.5 12.3 9.8 6.2 3.8t
FB 4 Benign 4+ 3.9 34 9.4 8.2 5.3 4.7
Baseline PSA 4.01-10 ng./ml.
GLa 5 Prostate Ca 3 4.4 22.7 11.5 7.0 5.6t
RM 6 Benign 4 4.4 23.8 34.3 22.2 12 4.4t
GT 6 Benign 3+ 4.5 25.0 10.2 7.1 6.1
DM 6 Benign 3 4.7 39.5 8.5 6.8 4.8t
JE 4 Prostate Ca 4 4.8 13 6.9 7.0 10 4.9t
AL 4 Prostate Ca 3 4.9 13 7.2 7.0 5.0t
JG 4 Benign 5+ 5.0 20.4 19.1 20.3 12.1 9.5 7.2
RSi 5 Prostate Ca 3 5.3 30.1 10.1 6.9 4.lt
AP 4 Benign 4 5.4 22.1 7.5 11.5 8.5 6.lt
WH 4 Prostate Ca 3 5.8 24.8 9.2 7.9 6.3t
FV 4 Benign 3 5.9 33.0 93.4 12.0 6.6t
RSt 3 Prostate Ca 4 5.9 47.9 21.2 9.0 7.8 6.2t
WK 4 Benign 3 6.1 14.4 18 10.9 6.3t
SSig 5 Benign 3+ 6.6 41.6 12.5 8.7 8.0
GF 6 Benign 3 6.8 44 13.5 11.0 6.5t
cw 6 Benign 4+ 7.3 63.3 13.7 10.2 10.9 9.6
Baseline PSA greater than 10 ng./ml.
KJ 6 Benign 4 16.3 44 32.8 Not available 24.2 22.4t
* Week + indicates continuously elevated PSA at last available followup.
t Return to baseline.
 EFFECTS OF PROSTATIC MANIPULATIONS ON SERUM PROSTATE SPECIFIC ANTIGEN LEVELS 813
PSA
25~-----------------------------~

20

15

10

a~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

BIBX 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38

DAYS

--¾- ACTUAL -6 -- THEORETICAL

Elevation of serum PSA level following prostatic biopsy and delayed return to baseline level

TABLE 5. Summary of published reports on various prostate manipulations and serum PSA levels
Procedure Pt. Group No. Men Effect on PSA Reference
Rectal examination Prostate Ca/BPH/prostatitis 24 None Brawer et al 15
Ultrasonography Prostate Ca/BPH 64 None Hughes et al 17
Ultrasonography Prostatitis 21 1.3 X increase
Massage Urinary calculi 13 Minimum or no EI-Shirbiny et al 14
increase in 12
Massage BPH 16 1.9 ± 0.88 x increase Stamey et al 16
Massage + cystoscopy Unspecified 7 4.1 ± 2.2 X increase
Perinea! biopsy Unspecified 7 57 x increase
Transurethral resection of prostate BPH 8 53 X increase

difference among serum PSA levels taken immediately before diately following transperineal needle biopsy and recommended
and those taken 5 or 90 minutes after rectal examinations. postponing PSA measurement for 2 weeks following the pro-
Prostate massage causes an appreciable increase of serum cedure.18 We found that 92 of 100 men (92%) had elevated
PSA levels in a small proportion of patients. Stamey et al found serum PSA levels following biopsy and in 29 of these 92 men
that serum PSA levels (Yang, Pros-Check) increased nearly 2- (32%) serum PSA levels failed to return to baseline within 2
fold in 16 men immediately following prostate massage. 16 El- weeks, as would be expected according to the published serum
Shirbiny et al found minimal or no increase of serum PSA in PSA half-life of 2 to 3 days. Clearance of PSA from the plasma
12 of 13 men, while the remaining man had a 5.5-fold increase is unpredictable in patients with a traumatized prostate and
(1.7 to 9.3 ng./ml., Tandem-R) when measured 60 minutes PSA clearance does not always follow the decline slope estab-
following prostatic massage, 14 Our experience suggests that lished after total prostatectomy. Biopsy causes anatomical dis-
prostatic massage has a more pronounced effect on serum PSA ruption that leads to continuous leakage of PSA into prostatic
level than rectal examination; however, we observed a clinically stroma and circulation, which results in a post-traumatic in-
significant elevation in only 1 of 20 men following massage. flammatory response within the prostate, These are 2 possible
We were unable to verify the nearly 2-fold increase in serum explanations for our observations. Vv e encountered a greater
PSA levels found by Stamey et al1 3 following prostatic massage than 50-fold increase of serum PSA levels that was reported
in any patient. The different PSA assays used (Tandem-R Stamey et al1 3 infrequently, which may be accounted for in a
versus Pros-Check) and the vigor of prostatic massage may number of ways, including we used a different method of biopsy
partially account for this difference. (transrectal versus transperineal) and different PSA assay, and
The effect of transrectal ultrasonography on serum PSA we took a different number of tissue cores from the patients.
levels paralleled that of prostate massage. In a study of 64 men To minimize over-interpretation of our data we selected the
with prostate cancer (treated and untreated) and BPH Hughes highest pre-biopsy serum PSA value as the baseline, we calcu-
et al found no significant difference between serum PSA levels lated the biological variation based on different baseline PSA
obtained immediately before and 30 minutes after transrectal ranges (normal, low and high titer elevation) and we used the
ultrasonography; however, in the same study following ultra- longest reported serum half-life of PSA. Our own calculations
sonography they observed a small (1.5 to 2.10 ng./ml., Tandem- based on serial PSA measurements in 10 patients following
R) but statistically significant increase of 1.3-fold in 21 men radical prostatectomy reveal a somewhat shorter serum half-
with prostatitis. 17 We found that 4 of 36 men (11%) had life of approximately 2 days.
elevations in serum PSA levels following ultrasonography but The desirability of obtaining serum PSA measurements be-
no effect was observed in the remaining 32 men. Therefore, our fore prostatic manipulation cannot be overemphasized, al-
findings support, at least in part, the previous observations of though this may not always be possible. Despite the fact that
Hughes et al. 17 we found no instances of a clinically significant increase in
To our knowledge no study has been conducted to examine serum PSA after rectal examination, 4 of 43 men (9%) had
the overall effects of prostate biopsy on serum PSA levels. sufficient increases that could possibly result in falsely abnor-
Stamey et al reported a 57-fold increase in serum PSA imme- mal serum PSA levels in patients with high normal baseline
814 YUAN AND ASSOCIATES

PSA levels (for example 3 to 4 ng./ml.). Falsely abnormal serum of prostate specific antigen in patients with prostate cancer. J.
PSA levels following prostatic massage (1 of 17, 6%) and Urol., 142: 1011, 1989.
ultrasonography (3 of 27, 11 %) pose more of a problem. One 5. Stamey, T. A., Kabalin, J. N., McNeal, J. E., Johnstone, I. M.,
should consider postponing PSA measurements for 1 week Freiha, F., Redwine, E. A. and Yang, N.: Prostate specific antigen
in the diagnosis and treatment of adenocarcinoma of the pros-
following prostatic massage, ultrasonography or rectal exami- tate. II. Radical prostatectomy treated patients. J. Urol., 141:
nation. Alternatively, one may draw the PSA level immediately 1076, 1989.
after rectal examination as long as one realizes that any bor- 6. Stamey, T. A., Kabalin, J. N. and Ferrari, M.: Prostate specific
derline elevation may be iatrogenic and require a repeat meas- antigen in the diagnosis and treatment of adenocarcinoma of the
urement. A waiting period of 4 weeks after prostatic biopsy prostate. III. Radiation treated patients. J. Urol., 141: 1084,
should be allowed before the PSA determination can be consid- 1989.
ered valid. Based on our results, the likelihood of a persistently 7. Stamey, T. A., Kabalin, J. N., Ferrari, M. and Yang, N.: Prostate
elevated PSA at 4 weeks is 7% while measurement 2 or 3 weeks specific antigen in the diagnosis and treatment of adenocarci-
following biopsy has a chance of 29% and 19%, respectively noma of the prostate. IV. Anti-androgen treated patients. J.
Urol., 141: 1088, 1989.
(table 4). 8. Lange, P. H., Ercole, C. J., Lightner, D. J., Fraley, E. E. and
Thus, premature sampling of spuriously elevated serum PSA V-€ssella, R.: The value of serum prostate specific antigen deter-
levels in patients with organ-confined tumors may have re- minations before and after radical prostatectomy. J. Urol., 141:
sulted in erroneous conclusion as to the nature and usefulness 873, 1989.
of serum PSA levels as a staging tool. This may partially 9. Oesterling, J. E., Chan, D. W., Epstein, J. I., Kimball, A. W., Jr.,
account for the frequent overlap of serum PSA levels that was Bruzek, D. J., Rock, R. C., Brendler, C. B. and Walsh, P. C.:
previously reported in patients with these organ-confined tu- Prostate specific antigen in the preoperative and postoperative
mors and in patients with extracapsular extension; it may also evaluation of localized prostatic cancer treated with radical pros-
be the reason for the reputed inaccuracy of serum PSA as an tatectomy. J. Urol., 139: 766, 1988.
10. Partin, A. W., Carter, H.B., Chan, D. W., Epstein, J. I., Oesterling,
indicator of pathological tumor stages. Pre-biopsy serum PSA J. E., Rock, R. C., Weber, J. P. and Walsh, P. C.: Prostate
levels may be more useful in preoperative staging than previ- specific antigen in the staging of localized prostate cancer: influ-
ously believed. In all future reports investigators should con- ence of tumor differentiation, tumor volume and benign hyper-
sider the timing of the PSA levels with respect to any prostatic plasia. J. Urol., 143: 747, 1990.
manipulations. 11. Catalona, W. J., Smith, D. S., Ratliff, T. L., Dodds, K. M., Coplen,
In conclusion, we have verified previous reports that digital D. E., Yuan, J. J. J., Petros, J. A. and Andriole, G. L.: Measure-
rectal examination causes minimum changes in serum PSA ment of prostate-specific antigen in serum as a screening test
levels, and that prostatic massage and ultrasonography may for prostate cancer. New Engl. J. Med., 324: 1156, 1991.
cause clinically significant elevations in approximately 10% of 12. Carlton, C. E., Jr. and Scardino, P. T.: Initial evaluation, including
history, physical examination, and urinalysis. In: Campbell's
individuals. More importantly, we report that needle biopsy of Urology, 5th ed. Edited by P. C. Walsh, R. F. Gittes, A. D.
the prostate may elevate PSA levels for longer intervals than Perlmutter and T. A. Stamey. Philadelphia: W. B. Saunders Co.,
predicted by published half-life determinations based on meas- pp. 276-310, 1986.
urements obtained following total prostatectomy. Our data 13. Schifman, R. B., Ahmann, F. R., Elvick, A., Ahmann, M., Coulis,
suggest that the half-life of PSA clearance is variable and K. and Brawer, M. K.: Analytical and physiological characteris-
unpredictable in patients with an intact prostate gland that tics of prostate-specific antigen and prostatic acid phosphatase
has been traumatized by needle biopsy. in serum compared. Clin. Chem., 33: 2086, 1987.
14. El-Shirbiny, A. M., Nilson, T. and Pawar, H. N.: Serum prostate
Linda Black, Cindy Heidorn, Mary Lynn and Cindy specific antigen: hourly change/24 hours compared with prostatic
Gebhardt obtained the PSA samples and Tracy Lewis per- acid phosphatase. Urology, 35: 88, 1990.
formed the PSA assay. 15. Brawer, M. K., Schifman, R. B., Ahmann, F. R., Ahmann, M. E.
and Coulis, K. M.: The effect of digital rectal examination on
REFERENCES serum levels ofprostatic-specific antigen. Arch. Path. Lab. Med.,
112: 1110, 1988.
1. Wang, M. C., Valenzuela, L.A., Murphy, G. P. and Chu, T. M.: 16. Stamey, T. A., Yang, N., Hay, A. R., McNeal, J.E., Freiha, F. S.
Purification of a human prostate specific antigen. Invest. Urol., and Redwine, E.: Prostate-specific antigen as a serum marker
17: 159, 1979. for adenocarcinoma of the prostate. New Engl. J. Med., 317:
2. Lilja, H.: A kallikrein-like serine protease in prostatic fluid cleaves 909, 1987.
the predominant seminal vesicle protein. J. Clin. Invest., 76: 17. Hughes, H. R., Penney, M. D., Ryan, P. G. and Peeling, W. G.:
1899, 1985. Serum prostatic specific antigen: in vitro stability and the effect
3. Watt, K. W. K., Lee, P. J., M'Timkulu, T., Chan, W. P. and Loor, of ultrasound rectal examination in vivo. Ann. Clin. Biochem.,
R.: Human prostate-specific antigen: structural and functional suppl. 1, 24: 206, 1987.
similarity with serine proteases. Proc. Natl. Acad. Sci., 88: 3166, 18. Stamey, T. A.: Prostate specific antigen in the diagnosis and
1986. treatment of adenocarcinoma of the prostate. Monogr. Urol.,
4. Hudson, M. A., Bahnson, R. R. and Catalona, W. J.: Clinical use 10: 50, 1989.