2012 EDUCATIONAL BOOK - American Society of Clinical Oncology

The Case for Prostate Cancer Risk Reduction
by 5-Alpha Reductase Inhibitors
By Youssef S. Tanagho, MD, MPH, and Gerald L. Andriole, MD
Overview: Prostate cancer remains a significant public
health problem. The current approach with prostate-specific
antigen (PSA)-based screening has questionable effects on
prostate cancer–specific mortality but is clearly associated
with overdiagnosis of prostate cancer, especially relatively
low-risk and low-volume tumors. Methods to decrease overdiagnosis
include alterations in screening practices and, potentially,
the use of 5-alpha reductase inhibitors. This article
reviews the major trials that have evaluated 5-alpha reductase
inhibitors in this setting: the Prostate Cancer Prevention Trial
(PCPT) and the Reduction by Dutasteride Prostate Cancer
Events Trial (REDUCE). Although these trials enrolled different
PROSTATE CANCER risk reduction is a potentially
valuable strategy, as current approaches with PSAbased
screening, which have a potentially small effect on
prostate cancer mortality, are associated with a significant
risk of overdiagnosis of prostate cancer. 1-3 Most men diagnosed
with PSA-detected tumors in the United States receive
aggressive treatments. 4,5 These treatments often have
significant side effects and are costly in both human and
economic terms. Therefore, one plausible strategy to improve
the efficacy of PSA-based screening is to consider the
use of 5-alpha reductase inhibitors, as they may reduce
overdiagnosis of prostate cancer and improve the utility of
PSA as a marker for aggressive disease, 6-9 as well as reduce
adverse sequelae of BPH.
Overdiagnosis of Prostate Cancer
Welch and Black 10 identified three factors that result in
overdiagnosis of prostate cancer: 1) the existence of a silent
disease reservoir; 2) the use of activities (such as screening)
that lead to its detection; and 3) tumors with a long natural
history and, hence, limited cancer-specific mortality. Prostate
cancer may represent a “textbook” disease for overdiagnosis.
Multiple studies have shown a high prevalence of this
disease on autopsy, ranging from approximately 30% of men
in their 30s to up to 80% of men in their 80s. Thus, there is
a large silent reservoir of this disease that could be clinically
detected. 11-12 Second, it has been estimated that close to
70% of American men have undergone one or more PSA
tests, and evidence suggests that older men, those in their
70s and 80s, who stand to benefit the least from PSA-based
screening represent the age stratum most likely to be
screened. 13 Moreover, use of relatively low PSA thresholds
as a guide to select patients for biopsy could result in the
diagnosis of prostate cancer in a high proportion of men in
their 60s to 80s. 14 It has been estimated that 6 to 10 times as
many men are considered to have an abnormal PSA than are
destined to die of prostate cancer. Finally, most cases of
prostate cancer have a long natural history and a limited
cancer-specific mortality. Albertsen and colleagues 15 looked
at survival among men with localized prostate cancer. Men
with Gleason score 8 to 10 disease and no comorbidities had
more than twice the chance of dying of other causes than of
prostate cancer within 10 years. Men who had one or more
significant comorbidities and Gleason score 8 to 10 disease
92
patient populations, their findings are complementary and
suggest a potential role for these agents in prostate cancer
risk reduction. Use of 5-alpha reductase inhibitors results in
an approximate 25% reduction in the detection of prostate
cancer, reduces diagnosis of high-grade prostatic intraepithelial
neoplasia (PIN), and improves benign prostatic hyperplasia
(BPH)-related outcomes and the performance of PSA as a
diagnostic test for aggressive prostate cancer. Side effects
occur in a small percentage of men and consist of decreased
sexual function and libido as well as gynecomastia. The risk of
high-grade tumor development while receiving these agents is
uncertain.
were about 5 times as likely to have a nonprostate cancer–
related death. A contemporary clinical trial, the Göteborg
screening study, demonstrated that in a screened population
of about 20,000 Scandinavian men—a population demographically
predisposed to a relatively high overall death
rate from prostate cancer—only 122 died of prostate cancer
and more than 3,800 died of other causes at 14 years of
follow-up. 3 Similarly, of 367 men with screen-detected, localized
prostate cancers who were considered candidates for
radical prostatectomy in the observation arm of the U.S.
PIVOT trial, 31 died of prostate cancer, whereas 152 died of
other causes at a median 10-year follow-up. 16
Estimates of the magnitude of overdiagnosis of prostate
cancer can be made by comparing the screened to the “usual
care” arms of randomized screening trials. The PLCO trial,
which enrolled men across 10 participating U.S. centers,
showed that the screened arm had a 17% (95% CI: 11–22)
increased chance of diagnosis of prostate cancer compared
with the usual care arm at 7 to 10 years of follow-up. 4
However, it is known that the usual care arm for PLCO was
heavily contaminated with PSA testing. If one were to
compare the usual care arm of PLCO to a matched contemporary
population in the United States, the usual care arm
would have a 22% excess prostate cancer detection rate.
Furthermore, men in the usual care arm of PLCO had more
than twice as many prostate cancers discovered as would
have been otherwise anticipated for a similar group of men
in the pre-PSA era. 17 In ERSPC, a multicenter European
trial, there was a 71% increase in the detection of prostate
cancer in the screened arm compared with the usual care
arm, and a 64% increase was noted in the Göteborg trial. In
most European sites, screening was performed at a 4-year
interval, whereas in Scandinavia, it generally occurred at a
2-year interval. One illustration of the effect of overdiagnosis
of prostate cancer was offered by Caroll and colleagues, 18
who used the Göteborg data to demonstrate that if 1,000
From the Division of Urology, Washington University School of Medicine, St. Louis, MO.
Authors’ disclosures of potential conflicts of interest are found at the end of this article.
Address reprint requests to Gerald L. Andriole, MD, Division of Urologic Surgery,
Washington University in St. Louis, 4960 Children’s Place, Campus Box 8242, St. Louis,
MO 63110; email: andrioleg@wustl.edu.
© 2012 by American Society of Clinical Oncology.
1092-9118/10/1-10