use of 5 -reductase inhibitors for prostate cancer

Use of 5�-Reductase Inhibitors for Prostate CancerChemoprevention: American Society of Clinical Oncology/

American Urological Association 2008 Clinical Practice Guideline

Barnett S. Kramer, Karen L. Hagerty, Stewart Justman, Mark R. Somerfield,Peter C. Albertsen,* William J. Blot, H. Ballentine Carter, Joseph P. Costantino,

Jonathan I. Epstein, Paul A. Godley,† Russell P. Harris, Timothy J. Wilt,Janet Wittes,‡ Robin Zon and Paul Schellhammer§

From the National Institutes of Health, Bethesda, MD; American Society of Clinical Oncology, Alexandria, VA; University of Montana LiberalStudies, Missoula, MT; University of Connecticut Health Center, Farmington, CT; International Epidemiology Institute, Rockville, MD; Johns

Hopkins University, Baltimore, MD; National Surgical Adjuvant Breast and Bowel Project, Pittsburgh, PA; University of North Carolina atChapel Hill, Chapel Hill, NC; University of Minnesota School of Medicine, Minneapolis, MN; Statistics Collaborative,Washington, DC;

Michiana Hematology Oncology, Granger, IN; and Eastern Virginia Medical School, Norfolk, VA

Approved by the American Society of ClinicalOncology Board on July 21, 2008, and by theAmerican Urological Association Board on Octo-ber 17, 2008.

This document is being reprinted as sub-mitted without editorial or independent peerreview by the Editors of The Journal ofUrology®.

This guideline was developed through a col-laboration between the American Society of Clin-ical Oncology and the American Urological Asso-ciation, and has been published jointly byinvitation and consent in the Journal of ClinicalOncology and The Journal of Urology®. No partof this document may be reproduced or transmit-ted in any form or by any means, electronic ormechanical, including photocopy, recording orany information storage and retrieval system,without written permission by the American So-ciety of Clinical Oncology or the American Uro-logical Association.

* Financial interest and/or other relationshipwith GlaxoSmithKline.

† Financial interest and/or other relationshipwith GlaxoSmithKline.

‡ Financial interest and/or other relationshipwith Merck.

§ Financial interest and/or other relationshipwith Theralogix, Dendreon Corporation, South-western Oncology Group and ContraVac.

See Editorial on page 000.

Purpose: To develop an evidence-based guideline on the use of 5-�-reductaseinhibitors (5-ARIs) for prostate cancer chemoprevention.Methods: The American Society of Clinical Oncology (ASCO) Health ServicesCommittee (HSC), ASCO Cancer Prevention Committee, and the American Uro-logical Association Practice Guidelines Committee jointly convened a Panel ofexperts, who used the results from a systematic review of the literature to developevidence-based recommendations on the use of 5-ARIs for prostate cancer che-moprevention.Results: The systematic review completed for this guideline identified 15 ran-domized clinical trials that met the inclusion criteria, nine of which reportedprostate cancer period prevalence.Conclusion: Asymptomatic men with a prostate-specific antigen (PSA) �3.0ng/mL who are regularly screened with PSA or are anticipating undergoingannual PSA screening for early detection of prostate cancer may benefit from adiscussion of both the benefits of 5-ARIs for 7 years for the prevention of prostatecancer and the potential risks (including the possibility of high-grade prostatecancer). Men who are taking 5-ARIs for benign conditions such as lower urinarytract [obstructive] symptoms (LUTS) may benefit from a similar discussion,understanding that the improvement of LUTS relief should be weighed with thepotential risks of high-grade prostate cancer from 5-ARIs (although the majorityof the Panel members judged the latter risk to be unlikely). A reduction ofapproximately 50% in PSA by 12 months is expected in men taking a 5-ARI;however, because these changes in PSA may vary across men, and within indi-vidual men over time, the Panel cannot recommend a specific cut point to triggera biopsy for men taking a 5-ARI. No specific cut point or change in PSA has beenprospectively validated in men taking a 5-ARI.

1642 www.jurology.com

INTRODUCTIONSurveys1,2 of American Society ofClinical Oncology (ASCO) members
have shown strong interest in cancer
0022-5347/09/1814-1642/0THE JOURNAL OF UROLOGY®

Copyright © 2009 by AMERICAN UROLOGICAL ASSOCIATION

prevention interventions applicableto clinical practice, and most respon-dents envision increased utilization of
prevention in their practices. Chemo-
Vol. 181, 1642-1657, April 2009Printed in U.S.A.

DOI:10.1016/j.juro.2009.01.071

5�-REDUCTASE INHIBITORS GUIDELINE 1643

prevention, a strong area of research, is particularlyrelevant to practicing oncologists and urologists be-cause its application is well suited to the clinicalsetting in which shared decision making takes placebetween health professionals and individual pa-tients. To date, the strongest evidence of efficacy inthe field of chemoprevention has come from the hor-monally responsive tumors: breast cancer with ta-moxifen and raloxifene3,4 and prostate cancer withthe 5-�-reductase inhibitor (5-ARI).5 Unlike mostother chemopreventive agents under study, such ascyclooxygenase-2 inhibitors and retinoids, tamox-ifen and finasteride have been shown definitively inrandomized clinical trials to decrease the incidenceof invasive cancers—not just surrogate endpoints—in healthy people.3–6 Nevertheless, theseagents have adverse effects that require careful dis-cussion with patients considering whether or not totake the agent.

Although substantial data are available on theuse of 5-ARIs in other settings—primarily, treat-ment of benign prostatic hyperplasia (BPH)7–10—the Prostate Cancer Prevention Trial (PCPT) is theonly completed randomized trial prospectively de-signed to show a reduction in period-prevalence ofprostate cancer.5 The PCPT investigators reported adecrease in cumulative incidence of prostate cancerfrom 24.4% in the placebo arm to 18.4% in the fin-asteride arm during the 7 years of the trial. Never-theless, an observed increase of Gleason scores 7 to10 in the finasteride study arm (37%, or 280 of 757tumors) compared with the placebo arm (22.2%, or237 of 1,068 tumors), noted in a secondary analysis,triggered concern about harm.

These issues complicate informed decision mak-ing by patients who are considering finasteride forprostate cancer chemoprevention, and are impor-tant to the many men taking finasteride for themanagement of BPH or for male pattern baldness.Published proposals discuss the balance of risks andbenefits of finasteride for the prevention of prostatecancer,11 although some have questioned the as-sumptions used in the calculation as overly favor-able.12 Because of the importance of the issue to bothclinical oncologists and urologists, ASCO has collab-orated with the American Urological Association(AUA) to develop a clinical practice guideline on thebenefits and harms of 5-ARIs for the prevention ofprostate cancer. New information will likely becomeavailable from ongoing analyses in the PCPT andthe results of the REDUCE (Reduction by Dutas-teride of Prostate Cancer Events) trial,13 a preven-tion trial testing dutasteride, which inhibits bothisoforms (types 1 and 2) of 5-�-reductase. This guide-line, sponsored by ASCO and AUA, aims to providea useful tool for clinicians and their patients in
making an informed decision about the potential
harms and benefits of taking 5-ARIs for preventingprostate cancer. As part of its deliberations, thePanel also evaluated the outcomes of the PLESS(Proscar Long-Term Efficacy and Safety Study) andMTOPS (Medical Therapy of Prostatic Symptoms)trials, which examined 5-ARIs for the relief of uri-nary tract obstruction, because reduced incidence ofurinary tract obstruction is one of the potential ben-efits of 5-ARI use. The charge to the Panel was tojudge the balance of benefits and harms.

GUIDELINE QUESTIONS

This guideline addresses the use of 5-ARIs in pre-venting prostate cancer. The overarching questionwas should men routinely be offered a 5-ARI for thechemoprevention of prostate cancer? To address thisquestion, the committee identified important compo-nents likely to influence assessments of the balanceof risks and benefits and decision making that re-quired evaluation:

1. What is the impact of 5-ARIs on the risk of inci-dent prostate cancer, prostate cancer mortality,and overall mortality? Do benefits and harms of5-ARIs vary among identifiable subpopulations(eg, age, race/ethnicity, family history, baselinerisk for prostate cancer) and by type of 5-ARI?

2. Do 5-ARIs have a differential effect on the devel-opment of different histologic grades or stages ofprostate cancer? Are any such differences likelyto modify the curability of prostate cancer whendiagnosed? Is the Gleason histologic grading sys-tem for prostate cancer applicable to men who arereceiving 5-ARIs or other interventions that tar-get the androgen pathway?

3. What is the impact of 5-ARIs on the need fortreatment of benign prostatic disease?

4. What is the impact of 5-ARIs on quality of life?What are other potential harms and adverse ef-fects of 5-ARIs? What are other potential benefitsof, and indications for, 5-ARI use (eg, benign pros-tatic hyperplasia, male baldness)?

5. How long should treatment continue for the bestoutcome (period v lifelong)?

6. What are the future directions of research re-garding 5-ARIs for the prevention of prostatecancer?

PRACTICE GUIDELINES

Practice guidelines are systematically developedstatements to assist practitioners and patients inmaking decisions about appropriate health care forspecific clinical circumstances. Attributes of goodguidelines include validity, reliability, reproducibil-
ity, clinical applicability, clinical flexibility, clarity,

5�-REDUCTASE INHIBITORS GUIDELINE1644

multidisciplinary process, review of evidence, anddocumentation. Guidelines may be useful in produc-ing better care and decreasing cost. Specifically, uti-lization of clinical guidelines may provide the follow-ing:

1. Improvement in outcomes2. Improvement in medical practice3. Means for minimizing inappropriate practice

variation4. Decision support tools for practitioners5. Points of reference for medical orientation and

education6. Criteria for self-evaluation7. Indicators and criteria for external quality re-

view8. Assistance with reimbursement and coverage

decisions9. Criteria for use in credentialing decisions

10. Identification of areas where further research isneeded.

In formulating recommendations for the use of5-ARIs for prostate cancer prevention, ASCO andAUA have considered these tenets of guideline de-velopment, emphasizing review of data from appro-priately conducted and analyzed clinical trials.These same tenets can be utilized and applied toformulation of prevention guidelines. However, it isimportant to note that guidelines cannot always ac-count for individual variation among patients.Guidelines are not intended to supplant physicianjudgment with respect to particular patients or spe-cial clinical situations and cannot be considered in-clusive of all proper methods of care or exclusive ofother treatments reasonably directed at obtainingthe same result.

ASCO and AUA’s practice guidelines and technol-ogy assessments reflect expert consensus based onclinical evidence and literature available at the timethey are written, and are intended to assist physi-cians in clinical decision making and identify ques-tions and settings for further research. Due to therapid flow of scientific information in oncology, newevidence may have emerged since the time a guide-line or assessment was submitted for publication.Guidelines and assessments are not continually up-dated and may not reflect the most recent evidence.Guidelines and assessments cannot account for in-dividual variation among patients, and cannot beconsidered inclusive of all proper methods of care orexclusive of other treatments. It is the responsibilityof the treating physician or other health care pro-vider, relying on independent experience and knowl-edge of the patient, to determine the best course oftreatment for the patient. Accordingly, adherence to
any guideline or assessment is voluntary, with the
ultimate determination regarding its application tobe made by the physician in light of each patient’sindividual circumstances. ASCO and AUA guide-lines and assessments describe the use of proceduresand therapies in clinical practice and cannot be as-sumed to apply to the use of these interventions inthe context of clinical trials. ASCO and AUA assumeno responsibility for any injury or damage to personsor property arising out of or related to any use ofASCO and AUA’s guidelines or assessments, or forany errors or omissions.

METHODS

Panel CompositionThe ASCO Health Services Committee (HSC) and theAUA Practice Guidelines Committee jointly convened anExpert Panel (hereafter referred to as the Panel) consist-ing of experts in clinical medicine and research methodsrelevant to chemoprevention for prostate cancer. The ex-perts’ fields included medical oncology, urology, pathology,epidemiology, statistics, and health services research. ThePanel included academic and community practitioners aswell as a patient representative. The Panel members arelisted in Appendix Table A1.

Literature Review and AnalysisWilt et al systematic review. The AUA commissioned asystematic review of the literature on the role of 5-ARIs inthe chemoprevention of prostate cancer. That review,which is published in the Cochrane Library,14 served asthe primary source of evidence for this guideline. Articleswere selected for inclusion in the systematic review if theymet the following prospective criteria: (1) a randomizedcontrolled trial (RCT) that examined a 5-ARI versus acontrol; (2) treatment period that was at least 1 year; (3)study population that consisted of men age 45 years orolder; (4) period prevalence of prostate cancer was one ofthe reported outcomes; and (5) the study was publishedafter 1984. For nonprostate cancer outcomes, randomizedtrials were selected for inclusion if they met the followingcriteria: (1) treatment period was at least 6 months; (2) anRCT compared a 5-ARI to a nonactive control or to anotheractive treatment; (3) study population consisted of menage 45 years or older; (4) at least one of the secondaryoutcomes of interest chosen by the Panel were reported;and (5) the study was published after 1999. This date waschosen because the AUA guideline on the management ofBPH included studies published through 1999.7 PCPTauthors provided unpublished information at the requestof the panel. Additional details of the literature searchstrategy and meta-analyses are provided elsewhere.14

ASCO/AUA Expert Panel literature review and analy-sis. The Panel reviewed all data from the primary studiescontained in the systematic review; subsequently, thePanel considered the results of the meta-analyses con-tained in the systematic review.

Consensus Development Based on EvidenceA steering committee met in October 2005. The steering
committee was charged with identifying potential Panel


members and with drafting the clinical questions thePanel was to address. In August 2006 the entire Panelmet; the Panel completed its additional work throughteleconferences. The purposes of the Panel meeting wereto refine the questions addressed by the guideline and tomake writing assignments for the respective sections. Allmembers of the Panel participated in the preparation ofthe draft guideline, which was then disseminated for re-view by the entire Panel. Feedback from external review-ers was also solicited. The ASCO HSC and the AUA Prac-tice Guidelines Committee, as well as the Board ofDirectors of both organizations, reviewed and approvedthe content of the guideline and the manuscript beforedissemination.

Guideline and Conflicts of InterestAll members of the Expert Panel complied with the ASCOand AUA policies on conflicts of interest, which requiredisclosure of any financial or other interest that might beconstrued as constituting an actual, potential, or apparentconflict. Members of the Expert Panel completed disclo-sure forms and were asked to identify ties to companiesdeveloping products that promulgation of the guidelinemight affect. Information was requested regarding em-ployment, consultancies, stock ownership, honoraria, re-search funding, expert testimony, and membership oncompany advisory committees. The Panel made decisionson a case-by-case basis as to whether an individual’s roleshould be limited as a result of a conflict. No limitingconflict was identified.

Revision DatesAt biannual intervals, the Panel Co-Chairs and two Panelmembers designated by the Co-Chairs will examine cur-rent literature to determine if there is a need for revisionsto the guideline. If necessary, the entire Panel will recon-vene to discuss potential changes. When appropriate, thePanel will recommend revision of the guideline to theASCO HSC, the ASCO Board, the AUA Practice Guide-lines Committee, and the AUA Board for review and ap-proval.

Definition of TermsPeriod prevalence: the proportion of the randomized pop-ulation identified as having prostate cancer over the trialperiod.

5-�-reductase: an enzyme that converts testosteroneinto the more potent dihydrotestosterone (DHT). 5-�-Re-ductase has two isoenzymes (isoforms): types 1 and 2.Finasteride inhibits the type 2 isoenzyme, whereas dutas-teride inhibits both the type 1 and type 2 isoenzymes.

Primary prevention: Intervention for relatively healthyindividuals with no invasive cancer and an average riskfor developing cancer.15

Chemoprevention: the use of chemical compounds toreduce the risk of development of a specific disease (asused here, specifically for the reduction in risk of develop-ing prostate cancer).

LUTS: Lower urinary tract [obstructive] symptoms.AUA Symptom Index/International Prostate Symptom

Score (IPSS): a symptom index for BPH, developed by theAUA. The IPSS consists of seven questions, with scores
ranging from 0 to 35. A score between 0 and 7 is consid-
ered to represent mild symptoms; 8 to 19, moderate; and20 to 35, severe. A change in score by 2 to 3 points isgenerally accepted as clinically meaningful (ie, morbidand/or life threatening).

Summary of Outcomes AssessedThe primary outcome assessed was either prostate cancerincidence or period prevalence, in prostate cancers de-tected “for cause.” For-cause cancers were defined as thosethat (1) were suspected clinically during the course of thetrial because of symptoms, abnormal digital rectal exam,or abnormal PSA (ie, a PSA that exceeded a certain valueor rate of increase over time) and were confirmed on bi-opsy; or (2) during the trial, a recommendation was madefor biopsy according to the study protocol (eg, due to in-creasing PSA) which was never done, and end-of-studybiopsy showed prostate cancer; or (3) end-of-study biopsyin the setting of a PSA more than 4 or a suspicious digitalrectal examination (DRE) showed prostate cancer.

Other primary outcomes assessed were distribution ofstage of prostate cancer, Gleason scores, and incidence orperiod prevalence of prostate cancer by age, race, baselinePSA, and family history.

Secondary outcomes included prostate cancer detectedpurely for reasons dictated by study protocol, rather thanby clinical indication (eg, an end-of-study biopsy was re-quired per protocol despite a PSA �4 ng/mL, and DRE notsuggestive of cancer), and overall incidence or period prev-alence of prostate cancer (ie, cancers detected for causeplus cancers detected by study protocol). Other secondaryoutcomes included quality of life; change in validated uri-nary symptom scale scores (IPSS/AUA); BPH progression(development of acute urinary retention, interventions fortreatment of LUTS, medications and herbal therapies totreat LUTS, surgical interventions); overall mortality;prostate cancer-specific mortality; adverse events (impo-tence, retrograde ejaculation, decreased ejaculate volume,decreased libido, gynecomastia); and cost effectiveness.Neither formal cost effectiveness nor decision analysiswas planned for the guideline, although the Panel consid-ered available reports.

RESULTS

Literature Search

Wilt et al14 identified 15 RCTs that met the inclu-sion criteria; nine of these trials reported prostatecancer period-prevalence.

Previous Guidelines and Consensus Statements

Updated in 2003,7 the 1999 AUA guideline on themanagement of benign prostate hyperplasia can besummarized as follows:

Finasteride, a 5-ARI, demonstrates both efficacyand acceptable safety for treatment of LUTS due toBPH. Finasteride can reduce the size of the prostate,can increase peak urinary flow rate, and can reduceBPH symptoms. With finasteride, the average pa-tient experiences a 3-point improvement in the AUASymptom Index. In general, patients perceive this
level of symptom improvement as a meaningful


change. Finasteride is ineffective for relief of lowerurinary tract obstructive symptoms in patients whodo not have enlarged prostates. Reported adverseevents are primarily sexually related; they includedecreased libido, ejaculatory dysfunction, and erec-tile dysfunction, which are reversible and thereforeuncommon after the first year of therapy. Symptomscore improvement is not substantially greateramong men with very large versus only moderatelyenlarged prostates; however, because of the moreprogressive nature of the disease in men with largerglands or higher PSA values, such patients conser-vatively treated (watchful or placebo groups) face anincreasingly higher risk of complication, thereby en-hancing the difference over time in outcomes be-tween finasteride and no treatment or placebogroups.

In summary, finasteride reduces the risk of sub-sequent acute urinary retention and the need forBPH-related surgery with the absolute benefit in-creasing with rising prostate volume or serum PSA.Finasteride is not appropriate treatment for menwith LUTS who do not have any prostatic enlarge-ment.

RECOMMENDATIONS

Should Men Routinely Be Offered a 5-ARI forthe Chemoprevention of Prostate Cancer?

Asymptomatic men with a PSA �3.0 who areregularly screened with PSA or are anticipating un-dergoing annual PSA screening for early detection ofprostate cancer may benefit from a discussion of thebenefits of 5-ARIs for 7 years for the prevention ofprostate cancer and the potential risks (includingthe possibility of high-grade prostate cancer) to beable to make a better-informed decision. Men whoare taking 5-ARIs for benign conditions such asLUTS would benefit from a similar discussion.

Of note, one Panel member believed that the ex-planation of decreased prostate cancer period prev-alence in the finasteride arm of the PCPT is attrib-utable to differential biopsy rates between the twotrial arms. The 25% risk reduction is based primar-ily on the PCPT and the risk reduction of a prostatediagnosis accrued primarily as a result of the lowerrates of biopsy among men on finasteride. For thosemen who underwent a biopsy, the risk reduction wasa statistically nonsignificant 10%. By contrast, theprincipal investigator of the PCPT has published acommentary stating that differential biopsy rate isnot likely to account for a substantial proportion ofthe observed difference. He argues that critics of thePCPT “. . . ignore the effects of PSA, DRE, and bi-opsy detecting the cancer in the finasteride group,which would be expected to further reduce the risk
of cancer overall.”
Evidence Summary

The summary of evidence of the effect of 5-ARIs onprostate cancer period prevalence is based on nineRCTs in which administration of the 5-ARI rangedfrom 1 to 7 years.14 Two additional trials lasting atleast 6 months provided data on BPH-related out-comes and adverse effects. Most of the trials wereplacebo controlled and double-blinded (AppendixTable A2). The results of all trials were generallyconsistent. The most salient end point chosen by thePanel was the comparison in trials of at least 1 yearduration of period prevalence of for-cause prostatecancers between study arms, as defined in the Meth-ods section and glossary, using an intent-to-treatanalysis (five studies). This end point comes closestto addressing prevention of those cancers that areconsidered most important in clinical decision mak-ing—those diagnosed because of symptoms, DREthat is clinically suggestive, or a PSA that triggers abiopsy. Only one of the randomized trials, the pla-cebo-controlled PCPT, was specifically designed toassess the period prevalence of prostate cancer ingenerally healthy men without underlying moderateor severe lower urinary track symptoms. The PCPTwas the largest trial in the literature, with 18,882 ofthe total 33,403 participants in trials receiving atleast 1 year of active therapy. The PCPT also con-tributed approximately 85% (1,006 of 1,188) of thefor-cause cancers in the systematic review.14 Men inall of the trials were screened regularly for prostatecancer with PSA and DRE. Therefore, the literaturereviewed does not allow assessment of the impact of5-ARIs on the period prevalence of prostate canceramong men who are not being actively screened forprostate cancer.

Review of Relevant Literature

What is the impact of 5-ARIs on the risk of incidentprostate cancer, prostate cancer mortality, and over-all mortality? Do benefits and harms of 5-ARIs varyamong identifiable subpopulations (eg, age, race/eth-nicity, family history, baseline risk for prostate can-cer) and by type of 5-ARI?

5-ARIs decrease the period prevalence of for-cause prostate cancer by approximately 26% (rela-tive risk � 0.74; 95% CI, 0.67 to 0.83; Fig 1). Theabsolute risk reduction is about 1.4% (4.9% in con-trols v 3.5% in 5-ARI arms), although this may varywith the age of the treated population. (For compar-ison, the data from the Breast Cancer PreventionTrial3 suggest that about 100 women with a 2%baseline risk of breast cancer would have to betreated for approximately 6 years with tamoxifen toprevent one case of invasive breast cancer.) The
relative risk of any prostate cancer with 5-ARI treat-

aborat


ment versus controls was 0.74 (95% CI, 0.56 to 0.98),for an absolute risk reduction of 2.9% (9.2% v 6.3%;Fig 2).

The PCPT is the most relevant of the trials con-sidered because its target population was men withat most mild lower urinary track symptoms who hada normal DRE and PSA less than 3 ng/mL at studyentry. This population is probably closest to a trueprimary prevention population. In the PCPT, therelative risk of for-cause prostate cancers amongthose receiving finasteride versus placebo was 0.76(95% CI, 0.68 to 0.86). This reduction corresponds tothe need to treat approximately 71 healthy men forapproximately 7 years to prevent one case of pros-tate cancer (95% CI, 50 to 100).

Only the PCPT reported subgroup results by race/ethnicity, age, and family history of prostate cancer.The data showed no apparent difference in efficacyof finasteride within any of these subgroups (Appen-

Study

Review: 5-alpha reductase inhibitors (5-ARIs) and prostate c

Comparison: 5-ARI versus placebo

Outcome: Prostate cancer detected “for cause”

5-ARI PlaceboN/nN/nyrogetacbus ro

01 Finasteride: Mid-term duration (1 to 2 years) FSG American 1992 3/595 1/300 FSG Int. 1993 5/495 3/255 PROSPECT 1996 303/6 013/3 Subtotal (95% CI) 1,400 858Total events: 11 (5-ARI), 10 (Placebo)Test for heterogeneity: χ2 = 0.78, df = 2, P = .68, I2 = 0%Test for overall effect: z = 0.67, P = .50

02 Finasteride: Long-term treatment duration (> 2 years) PCPT 2003 435/9,423 571/9,459 PLESS 1998 36/1,524 45/1,516Subtotal (95% CI) 10,947 10,975Total events: 471 (5-ARI), 616 (Placebo)Test for heterogeneity: χ2 = 0.03, df = 1, P = .86, I2 = 0%Test for overall effect: z = 4.45, P < .00001

03 Dutasteride: Long-term treatment duration (> 2 years) ARIA/ARIB 2004 27/2,167 55/2,158Subtotal (95% CI) 2,167 2,158Total events: 27 (5-ARI), 55 (Placebo)Test for heterogeneity: not applicableTest for overall effect: z = 3.07, P = .002

Total (95% CI) 14,514 13,991Total events: 509 (5-ARI), 681 (Placebo)Test for heterogeneity: χ2 = 4.31, df = 5, P = .51, I2 = 0%Test for overall effect: z = 5.17, P < .00001

Figure 1. Prostate cancer detected for cause. 5-ARI, 5-�-reductasProscar Safety Plus Efficacy Canadian Two-Year study; PCPT, ProSafety Study. Adapted with permission from the Cochrane Coll

dix Table A3), but estimates are necessarily impre-

cise, especially with regard to race/ethnicity, be-cause approximately 92% of the participants in thePCPT were non-Hispanic whites.

Importantly, the participants in all the trialswere being actively screened for prostate cancer.This study design issue affects the interpretation ofthe potential effect of 5-ARIs on the risk of develop-ing prostate cancer. For example, regular screeningwith PSA is known to double the incidence of diag-nosed prostate cancer.16 The relative reduction inrisk of being diagnosed with prostate cancer of ap-proximately 26% must be interpreted in this con-text, yielding a net increase in prostate cancer diag-noses of approximately 48% for the combinedstrategy of screening plus chemoprevention. Thestudies provide no information as to whether themagnitude of risk reduction for the diagnosis ofprostate cancer achieved by 5-ARIs would be thesame, or considerably less, in men who are not being

prevention

IC %59

1.51 (0.16 to 14.48) 0.86 (0.21 to 3.56)

)49.1 ot 21.0( 94.0 0.74 (0.30 to 1.79)

0.76 (0.68 to 0.86) 0.80 (0.52 to 1.23) 0.77 (0.68 to 0.86)

0.49 (0.31 to 0.77) 0.49 (0.31 to 0.77)

0.74 (0.67 to 0.83)

2 0.5 1 2 5

vors 5-ARI Favors Placebo

95% CIRR (fixed)

RR (fixed)

itor; RR, relative risk; FSG, Finasteride Study Group; PROSPECT,ancer Prevention Trial; PLESS, Proscar Long-Term Efficacy and

ion.

ancer

0.

Fa

e inhibstate C

actively screened for prostate cancer.

aborat


None of the trials was large enough to detectclinically important differences in either prostatecancer-specific mortality or overall mortality, and nodifference was noted. The summary relative risk ofprostate cancer-specific mortality from the random-ized trials was 1.00, with wide 95% CIs (0.29 to3.47). Likewise, the relative risk for all-cause mor-tality for 5-ARIs versus controls was 1.06 (95% CI,0.95 to 1.18). Absolute rates of prostate cancer mor-tality and overall mortality were 0.5% and 5.0%,respectively. Nevertheless, the Panel judged thateven if 5-ARI treatment never translates into re-duced overall or prostate cancer-specific mortality,reduction in risk of prostate cancer diagnosis withthe consequent morbidity of treatment is a clinicallybeneficial end point in and of itself.

Do 5-ARIs have a differential effect on the devel-opment of different histologic grades or stages of

01 Finasteride: Mid-term treatment duration (1 to 2 years) Cote 1998 8/27 1/25 FSG American 1992 3/595 1/300 FSG Int. 1993 552/3 594/5 PROSPECT 1996 3/310 6/303Subtotal (95% CI) 1,427 883Total events: 19 (5-ARI), 11 (Placebo)

Test for overall effect: z = 0.37, P = .71

02 Finasteride: Long-term treatment duration (> 2 years) PCPT 2003 803/9,423 1,147/9,459 PLESS 1998 72/1,524 77/1,516Subtotal (95% CI) 10,947 10,975Total events: 875 (5-ARI), 1,224 (Placebo)Test for heterogeneity: χ2 = 2.86, df = 1, P = .09, I2 = 65.1%Test for overall effect: z = 1.91, P = .057

03 Dutasteride: Long-term treatment duration (> 2 years) ARIA/ARIB 2004 27/2,167 55/2,158Subtotal (95% CI) 2,167 2,158Total events: 27 (5-ARI), 55 (Placebo)Test for heterogeneity: not applicableTest for overall effect: z = 3.07, P = .002

Total (95% CI) 14,541 14,016Total events: 921 (5-ARI), 1,290 (Placebo)Test for heterogeneity: χ2 = 11.51, df = 6, P = .07, I2 = 48.0%Test for overall effect: z = 2.09, P = .036

Fav

obecalPIRA-5ydutSN/nN/nyrogetacbus ro

Test for heterogeneity: χ2 = 5.17, df = 3, P = .16, I2 = 42.0%

Review: 5-alpha reductase inhibitors (5-ARIs) and prostate c

Comparison: 5-ARI versus placebo

Outcome: Prostate cancer detected overall

Figure 2. Prostate cancer detected overall. 5-ARI, 5-�-reductaseProscar Safety Plus Efficacy Canadian Two-Year study; PCPT, ProSafety Study. Adapted with permission from the Cochrane Coll

prostate cancer? Are any such differences likely to

modify the curability of prostate cancer when diag-nosed? Is the Gleason histologic grading system forprostate cancer applicable to men who are receiving5-ARIs or other interventions that target the andro-gen pathway?

The original publication and analysis of the PCPTscreening trial5 reported a statistically significantreduction in the 7-year period prevalence of prostatecancer when finasteride was compared with placebo.A secondary analysis showed an unexpected statis-tically significant greater number of high Gleasongrade cancers in the finasteride treatment arm. Forevery thousand men, finasteride reduced the num-ber of prostate cancers from 60 to 45; however, if theobservations represent a true finding, the number ofhigh-grade cancers would increase from 18 to 21 inthis cohort.14 The concern of paramount importancewas the potential causal relationship of finasteride

7.41 (1.00 to 55.09) 1.51 (0.16 to 14.48)

)65.3 ot 12.0( 68.0 0.49 (0.12 to 1.94) 1.23 (0.40 to 3.78)

0.70 (0.65 to 0.77) 0.93 (0.68 to 1.27) 0.78 (0.60 to 1.01)

0.49 (0.31 to 0.77) 0.49 (0.31 to 0.77)

0.74 (0.56 to 0.98)

0.2 0.5 1 2 5 10

ARI Favors Placebo

IC %59 95% CI

prevention

RR (random)RR (random)

or; RR, relative risk; FSG, Finasteride Study Group; PROSPECT,ancer Prevention Trial; PLESS, Proscar Long-Term Efficacy and

ion.

0.1

ors 5-

ancer

inhibitstate C

to high-grade cancer. The PCPT investigators, prac-


ticing physicians, and the Panel all struggled withthe question of whether the increase in high-gradecancers reflected a real risk of finasteride or merelyconfounding factors.

Several subsequent analyses of the data from thetrial have addressed this question. An initial con-cern centered about the morphologic and histopatho-logic alterations that might be attributed to a hor-monal agent or an agent, such as finasteride, whichalters the hormonal milieu. Might morphologicchanges caused by finasteride mimic high-gradecancer, and therefore artifactually produce a higherGleason reading? A detailed pathologic review elim-inated the questions of morphologic artifact as areason for the higher number of Gleason grade tu-mors in the finasteride arm.17

A second question that was addressed by thePCPT investigators was the degree of sampling er-ror induced by the established effect of 5-ARIs onreduction in prostate volume. Finasteride reducesprostate volume by approximately 25% to30%.5,9,10,18 The smaller the prostate volume, theless likely the sampling error using the PCPT’s pro-tocol-prescribed systematic sextant needle biopsies.Biopsies of the smaller prostates increase the likeli-hood of detection of prostate cancer of all grades. Iffinasteride is more effective in preventing low-gradetumors than high-grade tumors, it is possible toobserve a reduction in overall period prevalence ofcancers, a reduction in low- and moderate-gradecancers, but an absolute increase in high-grade can-cers as a consequence of less sampling error in thefinasteride arm. Indeed, modeling incorporatingprostate volume suggests that the increase in high-grade cancers seen in the PCPT may be nearly com-pletely explained by enhanced detection and not tu-mor transformation or induction, and would not beexpected to translate into an increase in prostatecancer mortality.18

A series of multivariable logistic regression mod-els was used to predict the incidence of high-gradecancer in the finasteride and placebo arms.19 Amodel that did not incorporate grading informationfrom radical prostatectomies showed a nonsignifi-cant 14% increase in high-grade cancer (P �.12) inthe finasteride group; however, incorporating infor-mation on grading led to a 27% lower (P � .02)bias-adjusted estimated rate of high-grade cancer inthe finasteride arm (6.0%) than in the placebo arm(8.2%). Although subject to the uncertainty of mod-eling assumptions outcomes, this analysis lends ad-ditional support against high-grade disease induc-tion by finasteride.

A third question of potential bias revolves aroundthe possibility of differences between the finasterideand placebo study arms of the PCPT in the operating
characteristics of PSA and DRE. Analysis by re-
ceiver operating characteristic curves demonstrateheightened sensitivity of both PSA and DRE in thefinasteride arm relative to the placebo arm, suggest-ing an enhanced suspicion for (and detection of)prostate cancer in the finasteride arm.20 It is reas-suring that despite the greater proportion of pa-tients with biopsy Gleason scores of �7 in the finas-teride group relative to placebo, patients from thetwo groups who subsequently were treated with rad-ical prostatectomy showed no difference in patho-logic stage, nodal involvement, or margin status.This observation, however, was possibly driven bydifferential choices in the decision to undergo radi-cal prostatectomy.

In summary, the Panel judged that plausible rea-sons could have led to a spurious increase in high-grade cancers. The Panel believed that it was un-likely for an agent to increase the incidence of high-grade tumors and simultaneously decrease theincidence of low-grade tumors. Furthermore, at thistime there is no known pathologic adjustment factorfor men diagnosed with prostate cancer while takinga 5-ARI versus men not taking a 5-ARI. Nor is thereinformation on the long-term outcomes for a givenhistologic grade for men taking a 5-ARI versus mennot having received a 5-ARI. Therefore, until thisinformation is known, decisions regarding the nat-ural history of the disease and decisions regardingtreatment interventions should be based on the his-tologic information obtained on biopsy, regardless of5-ARI status. Such an effect would be difficult toexplain on the basis of known biologic mechanisms.Nevertheless, none of the observations providesproof that finasteride would not increase the trueincidence of high-grade cancers. Therefore, menshould be fully apprised of the remaining uncer-tainty surrounding high-grade cancers with finas-teride.

Lastly, a detailed analysis of core biopsy speci-mens compared standard prognostic findings forcancers detected in the finasteride and placebo armsand made the following observations. For cancersassigned Gleason score �6, the mean number ofcores positive was 1.4 in the finasteride arm and1.55 in the placebo arm (P � .024); the mean per-centage of cores positive for Gleason score 7 was 31.2and 36.7 (P � .009); and for cancers assigned Glea-son score �8, bilateral core positivity was 28.6% inthe finasteride arm and 44.6% in the placebo arm(P � .047). These data suggest that men takingfinasteride had smaller, less aggressive tumors ver-sus men taking placebo.21

Apart from chemoprevention for prostate cancer,many men take a 5-ARI for indications such as lowerurinary tract obstructive symptoms or male patternbaldness. The observed increase in high-grade pros-
tate cancers in men receiving finasteride in the


PCPT, and the theoretical possibility of increasedrisk of prostate cancer mortality, should be dis-cussed with these men. The considerations men-tioned above provide some reassurance that the in-crease in diagnosis of high-grade tumors is morelikely to be due to artifact than to an actual increasein aggressive cancers. In such cases, observed bene-fits must be weighed against theoretical harms inmen who are being treated for symptomatic or both-ersome conditions.

What Is the Impact of 5-ARIs on the Need forTreatment for Benign Prostatic Disease?

5-ARIs are established, US Food and Drug Ad-ministration-approved, treatments for symptomaticBPH. The systematic review of trials in men withestablished BPH confirmed the efficacy of this classof drugs for this condition. To date, only the PCPTprovides information that directly addresses the im-pact of 5-ARIs in relatively asymptomatic men whodo not require treatment of urinary tract symptoms.In the PCPT, the incidence of acute urinary reten-tion was decreased by about one third in the finas-teride arm (RR � 0.67; 95% CI, 0.59 to 0.76; absoluterates, 6.3% v 4.2%). Consistent with this observa-tion, the incidence of transurethral resection of theprostate was 1.9% in the placebo arm and 1.0% inthe finasteride arm, a statistically significant de-crease in the risk for surgical interventions.

What is the impact of 5-ARIs on quality of life?What are other potential harms and adverse effects of5-ARIs? What are other potential benefits of, andindications for, 5-ARI use (eg, benign prostatic hy-perplasia, male baldness)?

Benefits of 5-ARI. The question of the impact of5-ARIs on global quality of life is difficult to answerbased on the available data. Studies of 5-ARIs havenot assessed global or general health-related qualityof life using traditional quality-of-life measures;rather, most studies have included measures of uri-nary symptoms (eg, urinary retention), sexual func-tioning (eg, erectile dysfunction), and/or endocrineeffects (eg, gynecomastia). Mid- and long-term stud-ies (6 months or beyond) with finasteride and dutas-teride individually and in meta-analysis of studiesthat included men with underlying LUTS demon-strated a reduction in risk in acute urinary retention(from 5.6% to 3.3%; absolute risk difference, 2.3%) aswell as a reduction in the need for surgical interven-tion (from 3.3% to 1.7%; absolute risk difference,1.6%). The largest benefits were observed in menwith a baseline PSA greater than 4 ng/mL (largerprostates). In addition, one RCT showed a statisti-cally significant reduction in hematuria with finas-teride compared with placebo.22

Adverse events. 5-ARIs are associated with a con-
sistently higher frequency of adverse events than
placebo, albeit of small absolute magnitude (Table1). Virtually all RCTs show a 2% to 4% increase inreported erectile dysfunction and gynecomastia, anddecreases in ejaculate volume and libido for thetreatment arm. When all studies, mid- and long-term, were combined, the overall discontinuation orloss to follow-up rates for both the 5-ARI and placeboarms were approximately 15%. The combined dis-continuation and loss to follow-up rate specificallysecondary to adverse events was approximately 6%to 7% in studies in both the 5-ARI and placebopatients. When the PCPT specifically evaluated ad-verse events, there were more adverse effects caus-ing temporary discontinuation in the finasteride arm(the investigators do not specifically define tempo-rary in this setting). Sexual function and endocrineeffects, which were more common in the 5-ARI arm(statistically significant), included decreased libido,decreased ejaculate volume, and gynecomastia. Thesexual dysfunction associated with finasteride de-creased over time, although it remained statisticallysignificant. The magnitude of effect was smallerthan natural sources of variability in the study pop-ulation23; on a scale of 0 to 100 (a higher scoreindicates more sexual dysfunction), the largest effectof finasteride on sexual functioning was a difference(mean) of 3.21 points, compared with a difference of1.26 points for each year of older age.

In trials comparing 5-ARIs to an alpha blocker forthe management of LUTS, discontinuation rateswere similar in the two groups. Dizziness and pos-tural hypertension were statistically more frequentamong patients receiving 5-� blocker therapy.10,24,25

5-ARIs effect on PSA. The decrease in PSA levels by5-ARIs must be taken into account when judging thesignificance of a PSA level. In the PCPT, finasteridelowered the PSA by 50% after 12 months of therapy,and therefore a multiplier of 2 was used as a crite-rion for biopsy. The effects of 5-ARIs on PSA before12 months are variable. In the PCPT, the decline at3 years was greater than 50% which was adjusted bychanging the 2 multiplier in the finasteride arm to2.3.

A single study has investigated the changes inPSA level caused by a 1-mg dose, as is used intreatment of male pattern baldness.26 For men age50 years and older, 1 mg of finasteride had an effectsimilar to 5 mg (50% decrease) at the 1-year fol-low-up date. Information beyond 1 year is notavailable.

Dutasteride inhibits both the type 1 and type 2isoforms of 5-� reductase and causes a greater andmore consistent decrease in DHT. The amount ofPSA suppression with long-term dutasteride treat-
ment has not been reported.


In conclusion, a consistently uniform scale multi-plier is currently unavailable because of inter- andintraindividual variability of PSA levels, laboratoryvariables, variable compliance of patients with re-gard to drug schedule, and a paucity of data relatingchange of PSA to duration of therapy. No specific cutpoint or change in PSA has been prospectively vali-dated in men taking a 5-ARI.

What is the treatment duration required for bestoutcome (period versus lifelong)?

No trial has directly compared different durationsof 5-ARIs for the prevention of prostate cancer. The

Table 1. Adverse effects

Subgroup/Reference

5-�-Reduc

No./Total N

Impotence/erectile dysfunctionMid-term treatment duration (1 to 2 years)

ARIA/ARIB8 13/1,605FSG-American47 25/595FSG-International48 22/495PREDICT24 13/264PROSPECT41 49/310VA COOP25 29/310

Long-term treatment duration (�2 years)PCPT5 6,349/9,423Total 6,500/13,00

Decreased/abnormal ejaculate volumeMid-term treatment duration (1 to 2 years)

FSG-American47 26/595VA COOP25 6/310

Long-term treatment duration (�2 years)PCPT5 5,690/9,423Finasteride Long-Term Efficacy and Safety Study Group36 23/1,524Total 5,745/11,85

Decreased libidoMid-term treatment duration (1 to 2 years)

ARIA/ARIB8 91/2,167FSG-American47 32/595PREDICT24 9/264PROSPECT41 31/310VA COOP25 14/310

Long-term treatment duration (�2 years)PCPT5 6,163/9,423Total 6,340/13,06

GynecomastiaMid-term treatment duration (1 to 2 years)

ARIA/ARIB8 50/2167Long-term treatment duration (�2 years)

PCPT5 426/9,423Total 476/11,59

IncontinenceLong-term treatment duration (�2 years)

MTOPS10 7/768PCPT5 183/9,423Total 190/10,19

Increased urinary frequency/urgencyPCPT5 1,214/9,423

Abbreviations: FSG, Finasteride Study Group; PREDICT, Prospective European DoxTwo-Year study; VA COOP, Veterans Affairs Cooperative Studies BPH Study GroupAdapted with permission from the Cochrane Collaboration.

PCPT was the only reported trial designed to test

the efficacy of a 5-ARI (finasteride) for preventingprostate cancer. It is also the most reliable trialfor directly comparing the benefits and harms offinasteride in the most likely target population forinterventions to prevent prostate cancer. Giventhat finasteride was administered for a planned 7years, the Panel judged that until additional in-formation is available, finasteride should be givenfor 7 years if used for primary prevention. Oneongoing trial13 compares dutasteride with placebofor preventing prostate cancer. Dutasteride is ad-ministered for 4 years in that trial, so the results

ibitor Placebo

Relative Risk 95% CI% No./Total No. %

0.81 14/1,555 0.90 0.90 0.42 to 1.914.2 5/300 1.7 2.52 0.97 to 6.524.4 1/255 0.4 11.33 1.54 to 83.604.9 9/269 3.3 1.47 0.64 to 3.38

15.8 19/303 6.2 2.52 1.52 to 4.189.4 14/305 4.6 2.04 1.10 to 3.78

67.4 5,816/9,457 61.5 1.10 1.07 to 1.1250.0 5,878/12,444 47.2 1.71 1.11 to 2.65

4.4 5/300 1.7 2.62 1.02 to 6.761.9 4/305 1.3 1.48 0.42 to 5.18

60.4 4,473/9,457 47.3 1.28 1.24 to 1.311.5 8/1,516 0.53 2.86 1.28 to 6.37

48.5 4,490/11,578 38.8 1.75 1.07 to 2.85

4.2 46/2,158 2.1 1.97 1.39 to 2.795.4 4/300 1.3 4.03 1.44 to 11.303.4 5/269 1.9 1.83 0.62 to 5.40

10.0 19/303 6.2 1.59 0.92 to 2.764.5 4/305 1.3 3.44 1.15 to 10.34

65.4 5,635/9,457 59.6 1.10 1.07 to 1.2248.5 5,713/12,792 44.7 1.83 1.19 to 2.81

2.3 16/2158 0.74 3.11 1.78 to 5.45

4.5 261/9,457 2.8 1.64 1.41 to 1.914.1 277/11,615 2.4 2.13 1.15 to 3.95

0.91 6/737 0.81 1.12 0.38 to 3.321.9 208/9,457 2.2 0.88 0.73 to 1.071.9 214/10,194 2.1 0.89 0.73 to 1.08

12.9 1,474/9,457 15.6 0.83 0.77 to 0.89

nd Combination Therapy trial; PROSPECT, Proscar Safety Plus Efficacy Canadianrostate Cancer Prevention Trial. MTOPS, Medical Therapy of Prostatic Symptoms.

tase Inh

o.

2

2

9

0

1

azosin a; PCPT, P

of the trial should help address the question of


whether a shorter duration of a 5-ARI preventsprostate cancer.

What are the future directions of research regard-ing 5-ARIs for the prevention of prostate cancer?

The goal of developing a chemopreventive agentthat can reduce the risk of prostate cancer has beenachieved. Nevertheless, despite the availability ofone large and well-designed trial (PCPT) and anongoing large trial (REDUCE) to test the efficacy offinasteride and dutasteride, respectively, for the pri-mary prevention of prostate cancer, important ques-tions and directions for research remain. The criticalquestion regarding the effect of 5-ARIs on prostatecancer morbidity and mortality is yet to be an-swered: is the observed increase in higher gradeprostate cancers in men receiving finasteride in thePCPT real or artifactual? Additional investigation isnecessary, and the results of REDUCE will aid inthis assessment.

The dose of finasteride currently being used totreat male pattern baldness is 1 mg per day ratherthan the 5 mg per day used in the PCPT. It isunknown if a 1-mg dose is as effective as 5 mg inreducing the risk of prostate cancer. If the lowerdose is just as effective, the balance of benefits andharms would be more favorable. All of this informa-tion would be critical to making more reliable esti-mates of cost effectiveness of 5-ARIs for preventingprostate cancer. It would also be important to knowwhether 5-ARIs reduce the incidence of clinicallydetected cancer among men who are not being ac-tively screened. This would be especially pertinentshould the ongoing randomized trials of prostatecancer screening (Prostate, Lung, Colorectal, Ovary[PLCO] Screening Study; European RandomizedStudy of Prostate Cancer [ERSPC]) show that pros-tate cancer screening does not decrease prostatecancer mortality or that the harms of screening out-weigh the benefits. In addition, as mentioned above,men taking a 5-ARI are expected to experience aroughly 50% reduction in PSA by 12 months; how-ever, because these changes in PSA may vary acrossmen, and within individual men over time, thePanel cannot recommend a specific cut point to trig-ger a biopsy for men taking a 5-ARI. This is animportant topic for future research.

Behavioral and communication research isneeded on the type and effectiveness of informationfor men considering the use of a 5-ARI for preven-tion of prostate cancer and for men who are alreadytaking a 5-ARI for other indications. More generally,as models incorporating molecular data evolve, risk-stratification models may be developed to tailor in-dividual preventive therapy. Such models may in-clude individuals with germline susceptibility at
heightened risk for developing prostate cancer, es-
pecially earlier onset disease or as a secondmalignancy.

PHYSICIAN-PATIENT COMMUNICATION

Although the decision to embark on a preventive5-ARI regimen belongs in the final analysis to theindividual, the physician has an important role. Forthe man who is considering the use of these agentsfor prostate cancer prevention, it is essential thatthe physician present the available data and high-light the remaining uncertainty.

5-ARIs do not eliminate the risk of developingprostate cancer; they reduce its clinical incidence.This distinction should be made clear. The advisabil-ity of the drug as a chemopreventive agent dependson an estimate of probabilities—of risk. As yet,there is no widely established decision model or aidthat can help a man of age 50 or 60 years determineif he is well advised to take a 5-ARI. Thompson etal27,28 have developed a prostate cancer risk calcu-lator based on risk equations generated from thePCPT data that may have clinical applications. Therisk calculator is applicable to men who are similarto those who participated in the PCPT: at least 50years of age, no prior prostate cancer diagnosis, andPSA and DRE results that are less than 1 year old.The calculator provides a preliminary assessment ofprostate cancer risk and the risk of high-grade pros-tate cancer if a prostate biopsy is performed, andwas recently validated in a more ethnically diverseand a younger population than that studied in thePCPT.29 The calculator is accessible at http://www.compass.fhcrc.org/edmnci/bin/calculator/main.asp?t�prostate&sub�disclaimer&v�prostate&m�&x�Prostate%20Cancer.

It is uncertain whether the findings of increasedhigh-grade cancers on the finasteride arm of thePCPT are actually an artifact of the methodology ofthe PCPT itself, as many believe. Until that risk isbetter elucidated, physicians should be explicitabout the uncertainties. In addition, it is importantto point out that the impact of 5-ARIs on prostatecancer mortality is unknown, given that the PCPTwas not designed to address this question. Never-theless, it is clear that a 5-ARI does convey benefitsto some men, as it decreases the overall incidence ofprostate cancer and also prevents urinary tract ob-struction in some men.

In communicating the present state of knowledgeabout the risks and benefits of 5-ARIs, it may behelpful to use the model of a cohort of 1,000 mentaking 5-ARIs for 7 years. Based on the systematicreview by Wilt et al,14 it is estimated that 5-ARIswill result in a reduction of 15 prostate cancer cases
in this group of 1,000 men, and in a possible increase
http://www.compass.fhcrc.org/edmnci/bin/calculator/main.asp?t=prostate%26sub=disclaimer%26v=prostate%26m=%26x=Prostate%20Cancer





of three cases of high-grade cancer after approxi-mately 7 years of therapy.

Men also should be informed that no informationon the long-term effects of 5-ARIs exists beyondapproximately 7 years, and that whether or not thereduction in prostate cancer incidence will translateinto a reduction in prostate cancer mortality orlonger life expectancy remains unknown.

Physicians should inform men who are consider-ing a 5-ARI about the incidence of sexual adverseeffects. Such adverse effects as lowered libido asso-ciated with a 5-ARI have been reported consistently,but they are also reversible. The benefits of a 5-ARIin reducing BPH are also appreciable and unequiv-ocal. It has been suggested that the ideal candidatefor a 5-ARI regimen would be an individual who is“concerned about the development of prostate can-cer, has a higher-than-average risk of the disease,has urinary symptoms that may be relieved by fin-asteride, and is not sexually active.”30

In summary, it is recommended that the physi-cian:

1. inform the man who is considering a 5-ARI thatthese agents reduce the incidence of prostate can-cer, and be sure to be clear that these agents donot reduce the risk of prostate cancer to zero;

2. discuss the elevated rate of high-grade cancerobserved in the PCPT and inform men of thepotential explanations;

3. make it known to men that no information on thelong-term effects of 5-ARIs on prostate cancerincidence exists beyond approximately 7 years,and that whether or not a 5-ARI reduces prostatecancer mortality or increases life expectancy re-mains unknown;

4. inform men of possible but reversible sexual ad-verse effects; and

5. inform men of the likely improvement in lowerurinary tract symptoms.

LIMITATIONS OF THE LITERATURE

Despite high-quality evidence from randomized trials,the data have limitations. Only one trial reported todate (PCPT) was specifically designed to measure theeffect of a 5-ARI on the incidence of prostate cancer; itwas not designed with sufficient power to assess theeffect of 5-ARIs on the risk of prostate cancer death. Todevelop a complete assessment of the benefit of5-ARIs, one would need to know the proportion ofcancers finasteride prevents that are truly clinicallymeaningful. The current literature cannot provide anestimate of this proportion; in particular, data arelacking on clinically meaningful cancers in the agegroup under consideration (men older than 50 years).
A recent analysis of PCPT characterized 34% of can-
cers as clinically insignificant by histologic criteria, arate similar to contemporary series of men who un-dergo treatment.21 Many cancers prevented by finas-teride might never have caused harm, as suggested bythe fact that screening leads to substantial overdiag-nosis of nonlethal cancers. Overdiagnosis can increasesubstantially with a lowering of the PSA threshold forprostatic biopsy or an increase in the number of sys-tematic biopsies of the prostate. Overdiagnosis islikely to decrease if ongoing randomized prostate can-cer screening trials show a net harm from screening.Therefore, the clinical importance of the prostate can-cers diagnosed in existing and ongoing chemopreven-tion trials is difficult to interpret. The Panel attemptedto focus on cancers that were most likely to be ofclinical importance (for-cause cancers as described inMethods). One of the most serious limitations in thecurrent literature is the changes that occur in stan-dards for PSA thresholds, criteria for biopsy, and bi-opsy methods. Continuing changes may alter all eval-uations of outcomes.

Although not specifically within the scope of thisdocument, the Panel did review studies on cost effec-tiveness. Two analyses31,32 have been published on thecost effectiveness of finasteride for prostate cancer pre-vention using the period prevalence observed in thePCPT and making a variety of unverifiable assump-tions about the impact of the diagnosed cancers onprostate cancer mortality and the tumor grade-specificlethality of the cancers detected. The results of thecost-effectiveness analyses were highly dependent onassumptions regarding whether the observed increasein high-grade tumors was real or artifactual, and onthe cost of drug. As extensively discussed above, thefirst assumption cannot be tested. Therefore, thePanel concluded that any assessment of cost effective-ness at this point is unreliable and impossible to in-corporate into the decision of whether or not to take5-ARIs for lowering the risk of prostate cancer.

SUMMARY OF DESIRED OUTCOMES

AND INTERPRETIVE SUMMARY

For the man who wishes periodic monitoring (oppor-tunistic or organized screening), 5-ARI therapy dur-ing a 7-year period reduces the period prevalence offor-cause cancer diagnoses by approximately 25%(relative risk reduction) for an absolute risk reduc-tion of about 1.4%. Although the majority of thePanel judged that the observed higher incidence ofhigh-grade (Gleason score 8 to 10) cancer in thefinasteride group is likely due to confounding fac-tors, the increased incidence of high-grade cancer asa result of induction by the drug cannot be excludedwith certainty. Additional benefits accruing from the
drug are reduction of the risk of urinary retention and


need for surgical intervention. Harms include sexualadverse effects, which usually diminish with time.

AUTHOR CONTRIBUTIONS

Conception and design: Barnett S. Kramer, TimothyJ. Wilt, Robin Zon, Paul Schellhammer

Administrative support: Karen L. Hagerty, MarkR. Somerfield

Collection and assembly of data: Barnett S.Kramer, Karen L. Hagerty, Mark R. Somerfield,Timothy J. Wilt, Paul Schellhammer

Data analysis and interpretation: Barnett S.Kramer, Karen L. Hagerty, Stewart Justman, PeterC. Albertsen, H. Ballentine Carter, Joseph P.Costantino, Jonathan I. Epstein, Paul A. Godley,Timothy J. Wilt, Janet Wittes, Paul Schellhammer

Manuscript writing: Barnett S. Kramer, Karen L.
Hagerty, Stewart Justman, Mark R. Somerfield, Pe-
ter C. Albertsen, Joseph P. Costantino, Jonathan I.Epstein, Timothy J. Wilt, Janet Wittes, Robin Zon,Paul Schellhammer

Final approval of manuscript: Barnett S. Kramer,Stewart Justman, Peter C. Albertsen, William J.Blot, H. Ballentine Carter, Joseph P. Costantino,Jonathan I. Epstein, Paul A. Godley, Russell P. Har-ris, Timothy J. Wilt, Janet Wittes, Robin Zon, PaulSchellhammer

ACKNOWLEDGMENTS

The Expert Panel thanks Erik Busby, MD, WilliamDahut, MD, Bernard Levin, MD, Ethan Basch, MD,J. Leonard Lichtenfeld, MD, Dean F. Bajorin, MD,Bruce J. Roth, MD, the ASCO Health Services Com-mittee and Board of Directors and the AUA PracticeGuidelines Committee, Board of Directors and staff,
Edie Budd, and Heddy Hubbard, PhD.
TABLE A15-�-Reductase Inhibitors Panel Members

Panel Member Institution

Barnett Kramer, MD, Co-Chair National Institutes of HealthPaul Schellhammer, MD, Co-Chair Eastern Virginia Medical SchoolStewart Justman, PhD, Patient Representative University of Montana Liberal StudiesPeter C. Albertsen, MD University of Connecticut Health CenterWilliam J. Blot, PhD International Epidemiology Institute, Vanderbilt-Ingram Cancer CenterH. Ballentine Carter, MD Johns Hopkins UniversityJoseph P. Costantino, PhD National Surgical Adjuvant Breast and Bowel ProjectJonathan I. Epstein, MD Johns Hopkins UniversityPaul A.Godley, MD, PhD University of North Carolina at Chapel HillRussell P. Harris, MD, MPH University of North Carolina at Chapel HillTimothy J. Wilt, MD, MPH Minneapolis VA Center for Chronic Disease Outcomes ResearchJanet Wittes, PhD Statistics CollaborativeRobin Zon, MD Michiana Hematology Oncology

TABLE A2Characteristics for 5-ARI Studies

Study (references) andInterventions (dose/d)

No. ofRandomlyAssignedPatients

TreatmentDuration(years) Baseline Demographic Characteristics

Long-term trials (�2 years)PCPT5 7 (� 90 days) American men enrolled onto a prostate cancer prevention trial

Finasteride, 5 mg 9,423 Age, 45–64 years, 62%; �65 years, 38%Placebo 9,459 Race/ethnicity: white, 92%; black, 3.8%; Hispanic, 2.6%; other, 1.5%

PSA (ng/mL), �3: 100%Family history of PCa (first-degree relative): 15.4%Mean baseline AUA/IPSS score: 6.7 (SD, 4.8)

MTOPS10,33 4.5 American men, mean age 62.6 (SD, 7.3), with LUTS secondary to BPHFinasteride, 5 mg 768 Race/ethnicity: white, 82.3%; black, 8.9%; Hispanic, 7.3%; other, 1.5%Doxazosin, 4–8 mg 756 Mean PSA (ng/mL): 2.4 (SD, 2.1)Combination finasteride and doxazsoin 786 Mean prostate volume (mL): 36.3 (SD, 20.1)Placebo 737 Mean baseline AUA/IPSS score: 16.9 (SD, 5.9)

PLESS9,34–37 4.0 American men, mean age 64 years (SD, 6.4), with LUTS secondary BPHFinasteride, 5 mg 1,524 Race: white 95.5%Placebo 1,516 Mean PSA (ng/mL): 2.8 (SD, 2.1).

Mean baseline AUA/IPSS score: 15 (SD, 5.7).History of sexual dysfunction: 46% of men in each group at screening

(appendix continued)


TABLE A2 (continued)

Study (references) andInterventions (dose/d)

No. ofRandomlyAssignedPatients

TreatmentDuration(years) Baseline Demographic Characteristics

Moderate-term trials (1–2 years)ARIA/ARIB 3001, 3002, 30038,38–40 2.0 Multinational men, mean age 66.3 years, with LUTS secondary BPH (all participants had IPSS �12)

Dutasteride, 0.5 mg 2,167 Race/ethnicity: white, 92%; black, 4%; Hispanic, 7.3%; Asian, 1%; other, 1.5%Placebo 2,158 Mean PSA (ng/mL): 4.0 (SD, 2.1)

Mean prostate volume (cm3): 54.5Mean baseline AUA/IPSS score: 17.05

PROSPECT41 2.0 Canadian men in good health, mean age, 63.3 years (range, 46 to 80 years) with LUTSsecondary to BPH

Finasteride, 5 mg 310 Race: NRPlacebo 303 Mean PSA (ng/mL): NR

Mean baseline AUA/IPSS score: 16.2PREDICT24 1.0 European men, mean age, 64 years (range, 50 to 80 years), with LUTS secondary to BPH

(all participants had IPSS �12)Finasteride, 5 mg 264 Race: NRDoxazosin, 4–8 mg 275 Mean PSA (ng/mL): 2.6Combination finasteride and doxazsoin 286 Mean prostate volume (g): 36Placebo 270 Mean baseline AUA/IPSS score: 17.2

VA Coop (Johnson, 2003; a subset ofsubjects from Lepor 1996)25,42

1.0 American men, mean age 65 years, with symptomatic BPH

Finasteride, 5 mg 252 Race: white 80%Terazosin, 10 mg 262 Race/ethnicity (from Lepor): white, 87%; black, 11%; Asian, 1%; Native American, 0.5%Combination finasteride and terazosin 272 Mean PSA (from Lepor; ng/mL): 2.3Placebo 254 Mean prostate volume (g): 37.6

Mean baseline AUA/IPSS score: 16.1Foley, 200022 1.0 British men, mean age, 77.5 years (range, 55 to 89 years), with hematuria associated with BPH

Finasteride, 5 mg 29 Race: NRWatchful waiting 28 Mean PSA (ng/mL): NR

Mean baseline AUA/IPSS score: NRCote, 199843 1.0 American men age �50 years (mean, 68 years) with elevated PSA (�4.0 ng/mL); study objective

was to examine effect of finasteride on prostate cellular proliferation and high-grade PINObservation (watchful waiting) 29 Race: NRFinasteride, 5 mg 29 Mean PSA (ng/mL): 9.8

Mean baseline AUA/IPSS score: NRPre-existing high-grade PIN: observation, 5 men; Finasteride, 8 men

FSG44–48 American and international men; mean age, 64.9 years (range, 40 to 83 years) with BPHFinasteride, 1 mg 547 Race/ethnicity: white, 95.8%; black, 1.6%; other, 2.6%.Finasteride, 5 mg 543 Mean PSA (ng/mL): 4.7Placebo 555 Mean prostate volume (cm3): 55.0

Mean baseline AUA/IPSS score: NRShort-term trials (1 year)

ARIA 200149 0.5 American and Canadian men age �50 years (mean, 62.6 to 65.5 years across groups), withLUTS secondary to BPH; study objective was to examine effect of dutasteride on serumdihydrotestosterone levels

Dutasteride, 0.01, 0.05, 0.5, 2.5, and5 mg

285 Race: NR

Finasteride, 5 mg 55 Mean PSA (ng/mL): NRPlacebo 59 Mean baseline AUA/IPSS score: NR

MICTUS50 0.5 Italian men, mean age, 63 years (SD, 7.1 years), with LUTS secondary to BPH (allparticipants had IPSS �13)

Finasteride, 5 mg 204 Race: NRTamsulosin, 0.4 mg 199 Mean PSA (ng/mL): NR

Mean baseline AUA/IPSS score: NRLee, 200251 Korean men, mean age, 64.7 years, with LUTS secondary to BPH (all participants had

Korean IPSS �8)Finasteride, 5 mg 102 Race: Asian, 100%Tamsulosin, 0.2 mg 103 Mean PSA (ng/mL): 2.0

Mean prostate volume (cm3): 29.8Mean baseline AUA/IPSS score: 19.5

Abbreviations: 5-ARI, 5-�-reductase inhibitor; PCPT, Prostate Cancer Prevention Trial; PSA, prostate-specific antigen; PCa, prostate cancer; AUA/IPSS, American UrologicalAssociation/International Prostate Symptom Score; SD, standard deviation; NR, not reported; MTOPS, Medical Therapy of Prostatic Symptoms; LUTS, lower urinary tractsymptoms; BPH, benign prostatic hyperplasia; PLESS, Proscar Long-Term Efficacy and Safety Study; PREDICT, Prospective European Doxazosin and Combination Therapy; PIN,
prostatic intraepithelial neoplasia; FSG, Finasteride Study Group; MICTUS, Multicentre Investigation to Characterise the Effect of Tamsulosin on Urinary Symptoms.


TABLE A3Period Prevalence of Prostate Cancer According to Subgroup

Subgroup/Reference

5-�-Reductase Inhibitor Placebo

RelativeRisk 95% CI

No./Total No.of Patients %

No./Total No.of Patients %

PSA at study entry: 0.0 to �4.0 ng/mLPCPT5 803/9,423 8.5 1,147/9,456 12.1 0.70 0.64 to 0.77PLESS36 28/1,149 2.4 32/1,154 2.8 0.88 0.53 to 1.45Total 831/10,572 7.9 1,179/10,610 11.1 0.71 0.65 to 0.77

PSA at study entry: �4.0 ng/mLCote43 8/27 29.6 1/25 4.0 7.41 1.00 to 55.09PLESS36 44/374 11.8 45/357 12.6 0.93 0.63 to 1.38Total 52/401 13.0 46/382 12.0 2.08 0.28 to 15.43

Gleason score 7PCPT5 190/9,423 2.0 184/9,459 1.9 1.04 0.85 to 1.27PLESS (estimated from graph)36 11/1,524 0.72 12/1,516 0.79 0.91 0.40 to 2.06Total 201/10,947 1.8 196/10,975 1.9 1.03 0.85 to 1.25

Gleason score 7: PCPT5

PCa overall/number randomized 190/9,423 2.0 184/9,459 1.9 1.04 0.85 to 1.27PCa overall/included in analysis 190/4,368 4.3 184/4,692 3.9 1.11 0.91 to 1.35PCa detected for cause/number randomized 118/9,423 12.5 103/9,459 10.9 1.15 0.88 to 1.50PCa detected for cause/biopsy performed for cause 118/1,639 7.2 103/1,934 5.3 1.35 1.05 to 1.75

Gleason scores 8 to 10PCPT5 90/9,423 0.96 53/9,459 0.56 1.70 1.22 to 2.39PLESS (estimated from graph)36 1/1,524 0.07 7/1,516 0.46 0.14 0.02 to 1.15

Gleason scores 8 to 10: PCPT5

PCa overall/No. randomly assigned 90/9,423 0.96 53/9,459 0.56 1.70 1.22 to 2.39PCa overall/No. included in analysis 90/4,368 2.1 53/4,692 1.1 1.82 1.30 to 2.55PCa detected for cause/No. randomly assigned 70/9,423 0.74 45/9,459 0.48 1.56 1.07 to 2.27PCa detected for cause/No. with biopsy performed

for cause70/1,639 4.3 45/1,934 2.3 1.84 1.27 to 2.65

Age, years: PCPT5

55 to 59 205/2,954 6.9 309/2,954 10.5 0.66 0.56 to 0.7960 to 64 254/2,970 8.6 357/2,825 12.6 0.68 0.58 to 0.79�65 344/3,498 9.8 481/3,677 13.1 0.75 0.66 to 0.86

Race/ethnicity: PCPT5

Non-Hispanic white 739/8,667 8.5 1067/8,713 12.2 0.70 0.64 to 0.76African-American 41/356 11.5 50/353 14.2 0.81 0.55 to 1.20Hispanic 19/262 7.3 23/237 9.7 0.75 0.42 to 1.34

Family history of PCa in first-degree relative: PCPT5

Yes 176/1,458 12.1 241/1,455 16.6 0.73 0.61 to 0.87No 627/7,965 7.9 906/8,002 11.3 0.70 0.63 to 0.77

Abbreviations: PSA, prostate-specific antigen; PCPT, Prostate Cancer Prevention Trial; PLESS, Proscar Long-Term Efficacy and Safety Study; PCa, prostate cancer.

REFERENCES

1. Chlebowski RT, Sayre J, Frank-Stromborg M, et al:Current attitudes and practice of American Societyof Clinical Oncology-member clinical oncologists re-garding cancer prevention and control. J Clin Oncol1992; 10: 164.

2. Ganz PA, Kwan L, Somerfield MR, et al: The role ofprevention in oncology practice: Results from a2004 survey of American Society of Clinical Oncol-ogy members. J Clin Oncol 2006; 24: 2948.

3. Fisher B, Costantino JP, Wickerham DL, et al: Ta-moxifen for prevention of breast cancer: Report ofthe National Surgical Adjuvant Breast and BowelProject P-1 Study. J Natl Cancer Inst 1998; 90: 1371.

4. Fisher B, Costantino JP, Wickerham DL, et al: Ta-
moxifen for the prevention of breast cancer: Current
status of the National Surgical Adjuvant Breast andBowel Project P-1 study. J Natl Cancer Inst 2005;97: 1652.

5. Thompson IM, Goodman PJ, Tangen CM, et al: Theinfluence of finasteride on the development of pros-tate cancer. N Engl J Med 2003; 349: 215.

6. Vogel VG, Costantino JP, Wickerham DL, et al:Effects of tamoxifen vs raloxifene on the risk ofdeveloping invasive breast cancer and other diseaseoutcomes: The NSABP Study of Tamoxifen andRaloxifene (STAR) P-2 trial. JAMA 2006; 295: 2727.

7. AUA guideline on management of benign prostatichyperplasia: (2003). Chapter 1: Diagnosis and treat-
ment recommendations. J Urol 2003; 170: 530.
8. Roehrborn CG, Boyle P, Nickel JC, et al: Efficacy andsafety of a dual inhibitor of 5-alpha-reductase types1 and 2 (dutasteride) in men with benign prostatichyperplasia. Urology 2002; 60: 434.

9. Kaplan S, Garvin D, Gilhooly P, et al: Impact ofbaseline symptom severity on future risk of benignprostatic hyperplasia-related outcomes and long-term response to finasteride: The Pless StudyGroup. Urology 2000; 56: 610.

10. McConnell JD, Roehrborn CG, Bautista OM, et al:The long-term effect of doxazosin, finasteride, andcombination therapy on the clinical progression ofbenign prostatic hyperplasia. N Engl J Med 2003;
349: 2387.


11. Klein EA, Tangen CM, Goodman PJ, et al: Assessingbenefit and risk in the prevention of prostate cancer:The prostate cancer prevention trial revisited. J ClinOncol 2005; 23: 7460.

12. Talcott JA: Rebalancing ratios and improving im-pressions: Later thoughts from the prostate cancerprevention trial investigators. J Clin Oncol 2005; 23:7388.

13. Andriole G, Bostwick D, Brawley O, et al: Chemo-prevention of prostate cancer in men at high risk:Rationale and design of the reduction by dutasterideof prostate cancer events (REDUCE) trial. J Urol2004; 172: 1314.

14. Wilt T, MacDonald R, Hagerty K et al: 5-Alpha-reductase inhibitors for prostate cancer prevention.Cochrane Database Sst Rev 2008; issue 2, art. No.CD007091. doi: 10.1002/14651858.CD007091.

15. Lippman SM, Levin B, Brenner DE, et al: Cancerprevention and the American Society of ClinicalOncology. J Clin Oncol 2004; 22: 3848.

16. Welch HG: Should I Be Tested for Cancer? MaybeNot and Here’s Why. Berkeley, CA, University ofCalifornia Press, pp 51–65.

17. Lucia MS, Epstein JI, Goodman PJ, et al: Finasterideand high-grade prostate cancer in the Prostate Can-cer Prevention Trial. J Natl Cancer Inst 2007; 99:1375.

18. Cohen YC, Liu KS, Heyden NL, et al: Detection biasdue to the effect of finasteride on prostate volume:A modeling approach for analysis of the ProstateCancer Prevention Trial. J Natl Cancer Inst 2007; 99:1366.

19. Redman MW, Tangen CM, Goodman PJ, et al:Finasteride does not increase the risk of high-grade prostate cancer: A bias-adjusted modelingapproach. Cancer Prev Res 2008; 1: 174.

20. Thompson IM, Chi C, Ankerst DP, et al: Effect offinasteride on the sensitivity of PSA for detectingprostate cancer. J Natl Cancer Inst 2006; 98: 1128.

21. Lucia MS, Darke AK, Goodman PJ, et al: Pathologiccharacteristics of cancers detected in the prostatecancer prevention trial: implications for prostatecancer detection and chemoprevention. Cancer PrevRes [epub ahead of print on May 18, 2008].

22. Foley SJ, Soloman LZ, Wedderburn AW, et al: Aprospective study of the natural history of hematuriaassociated with benign prostatic hyperplasia andthe effect of finasteride. J Urol 2000; 163: 496.

23. Moinpour CM, Darke AK, Donaldson GW, et al:Longitudinal analysis of sexual function reported bymen in the Prostate Cancer Prevention Trial. J NatlCancer Inst 2007; 99: 1025.

24. Kirby RS, Roehrborn C, Boyle P, et al: Efficacy andtolerability of doxazosin and finasteride, alone or incombination, in treatment of symptomatic benignprostatic hyperplasia: The Prospective EuropeanDoxazosin and Combination Therapy (PREDICT) trial.
Urology 2003; 61: 119.
25. Lepor H, Williford WO, Barry MJ, et al: The efficacyof terazosin, finasteride, or both in benign prostatichyperplasia: Veterans Affairs Cooperative StudiesBenign Prostatic Hyperplasia Study Group. N EnglJ Med 1996; 335: 533.

26. D’Amico AV, Roehrborn CG: Effect of 1 mg/dayfinasteride on concentrations of serum prostate-specific antigen in men with androgenic alopecia: Arandomised controlled trial. Lancet Oncol 2007; 8:21.

27. Thompson IM, Ankerst DP, Chi C, et al: Assessingprostate cancer risk: Results from the Prostate Can-cer Prevention Trial. J Natl Cancer Inst 2006; 98:529.

28. Thompson IM, Pauler Ankerst D, Chi C, et al: Pre-diction of prostate cancer for patients receivingfinasteride: Results from the Prostate Cancer Pre-vention Trial. J Clin Oncol 2007; 25: 3076.

29. Parekh DJ, Ankerst DP, Higgins BA, et al: Externalvalidation of the Prostate Cancer Prevention Trialrisk calculator in a screened population. Urology2006; 68: 1152.

30. Zuger A: A big study yields big questions. N EnglJ Med 2003; 349: 213.

31. Svatek RS, Lee JJ, Roehrborn CG, et al: The cost ofprostate cancer chemoprevention: A decision anal-ysis model. Cancer Epidemiol Biomarkers Prev 2006;15: 1485.

32. Zeliadt SB, Etzioni RD, Penson DF, et al: Lifetimeimplications and cost-effectiveness of using finas-teride to prevent prostate cancer. Am J Med 2005;118: 850.

33. Kaplan SA, McConnell JD, Roehrborn CG, et al:Combination therapy with doxazosin and finasteridefor benign prostatic hyperplasia in patients withlower urinary tract symptoms and a baseline totalprostate volume of 25 ml or greater. J Urol 2006;175: 217; discussion 220.

34. Paick SH, Meehan A, Lee M, et al: The relationshipamong lower urinary tract symptoms, prostate spe-cific antigen and erectile dysfunction in men withbenign prostatic hyperplasia: Results from theProscar Long-Term Efficacy and Safety Study. J Urol2005; 173: 903.

35. Wessells H, Roy J, Bannow J, et al: Incidence andseverity of sexual adverse experiences in finasterideand placebo-treated men with benign prostatic hy-perplasia. Urology 2003; 61: 579.

36. Andriole GL, Guess HA, Epstein JI, et al: Treatmentwith finasteride preserves usefulness of prostate-specific antigen in the detection of prostate cancer:Results of a randomized, double-blind, placebo-con-trolled clinical trial. PLESS Study Group—ProscarLong-Term Efficacy and Safety Study. Urology 1998;52: 195. discussion 201.

37. McConnell JD, Bruskewitz R, Walsh P, et al: Theeffect of finasteride on the risk of acute urinaryretention and the need for surgical treatment amongmen with benign prostatic hyperplasia: FinasterideLong-Term Efficacy and Safety Study Group. N Engl
J Med 1998; 338: 557.
38. Andriole GL, Roehrborn C, Schulman C, et al: Effectof dutasteride on the detection of prostate cancer inmen with benign prostatic hyperplasia. Urology2004; 64: 537; discussion 542.

39. Andriole GL, Kirby R: Safety and tolerability of thedual 5alpha-reductase inhibitor dutasteride in thetreatment of benign prostatic hyperplasia. Eur Urol2003; 44: 82.

40. O’Leary MP, Roehrborn C, Andriole G, et al: Im-provements in benign prostatic hyperplasia-specificquality of life with dutasteride, the novel dual 5al-pha-reductase inhibitor. BJU Int 2003; 92: 262.

41. Nickel JC, Fradet Y, Boake RC, et al: Efficacy and safetyof finasteride therapy for benign prostatic hyperplasia:Results of a 2-year randomized controlled trial (thePROSPECT study)—PROscar Safety Plus Efficacy Ca-nadian Two year Study. CMAJ 1996; 155: 1251.

42. Johnson TM II, Jones K, Williford WO, et al:Changes in nocturia from medical treatment of be-nign prostatic hyperplasia: Secondary analysis ofthe Department of Veterans Affairs CooperativeStudy Trial. J Urol 2003; 170: 145.

43. Cote RJ, Skinner EC, Salem CE, et al: The effect offinasteride on the prostate gland in men with ele-vated serum prostate-specific antigen levels. Br JCancer 1998; 78: 413.

44. Stoner E: Three-year safety and efficacy data on theuse of finasteride in the treatment of benign pros-tatic hyperplasia. Urology 1994; 43: 284; discussion292.

45. Stoner E: Maintenance of clinical efficacy with fin-asteride therapy for 24 months in patients withbenign prostatic hyperplasia: The Finasteride StudyGroup. Arch Intern Med 1994; 154: 83.

46. Stoner E, Round E, Ferguson D, et al: Clinical expe-rience of the detection of prostate cancer in patientswith benign prostate hyperplasia treated with fin-asteride: The Finasteride Study Group. J Urol 1994;151: 1296.

47. Gormley GJ, Stoner E, Bruskewitz RC, et al: Theeffect of finasteride in men with benign prostatichyperplasia: The Finasteride Study Group. N EnglJ Med 1992; 327: 1185.

48. Finasteride (MK-906) in the treatment of benignprostatic hyperplasia: The Finasteride Study Group.Prostate 1993; 22: 291.

49. Clark RV, Hermann DJ, Cunningham GR, et al:Marked suppression of dihydrotestosterone in menwith benign prostatic hyperplasia by dutasteride, adual 5alpha-reductase inhibitor. J Clin EndocrinolMetab 2004; 89: 2179.

50. Rigatti P, Brausi M, Scarpa RM, et al: A comparisonof the efficacy and tolerability of tamsulosin andfinasteride in patients with lower urinary tract symp-toms suggestive of benign prostatic hyperplasia.Prostate Cancer Prostatic Dis 2003; 6: 315.

51. Lee E: Comparison of tamsulosin and finasteride forlower urinary tract symptoms associated with be-nign prostatic hyperplasia in Korean patients. J Int
Med Res 2002; 30: 584.

use of 5 -reductase inhibitors for prostate cancer

Documents