After successful completion of the placebo run-in phase, subjects who continue to meet eligibility requirements will be randomized into the double-blind phase of the study and issued a 6-month supply of study drug. Subjects will self-administer study drug once daily orally for up to 4 years.

The occurrence and quantity of high-grade prostatic intraepithelial neoplasia (HGPIN) and atypical small acinar proliferation (ASAP) at biopsy were measured. HGPIN and ASAP are considered precancerous conditions. A participant diagnosed with prostate cancer only (i.e., no HGPIN or ASAP) was counted in both the first category ("HGPIN or prostate cancer diagnosis") and again in the last category ("HGPIN, ASAP, or prostate cancer diagnosis").

The amount of prostate biopsy tissue with HGPIN was measured.

The average amount of cancer seen by the pathologist in the prostate tissue samples taken during the biopsy was measured. A core is a prostate biopsy sample.

The average number of prostate biopsy samples (cores) determined to be cancerous by the pathologist was measured. Normally, 10 cores were taken per biopsy for each participant.

The treatment alteration score is a measure of the cellular changes due to treatment (effect of male hormone withdrawal) on the nucleus and cytoplasm of the prostate cancer cell. The treatment alteration score is the sum of two scores (the nuclear alteration score and the cytoplasmic architectural score), each ranging from 0 to 3, with 0 indicating no change and 3 indicating severe changes.

The number of participants who received treatment for prostate cancer was measured. Prostate cancer interventions included surgical interventions (e.g., prostatectomy, adenomectomy, transurethral resection) and non-surgical interventions (e.g., chemotherapy, hormone therapy, radiation therapy).

The IPSS is a 7-item questionnaire that measures urinary symptoms. It measures the level of urinary symptoms (including incomplete emptying, frequency, intermittency, urgency, weak stream, straining, and nocturia) reported as the total IPSS score. Each of the 7 questions has a 6-point response scale (0=none/not at all to 5=almost always) with a total score that can range from 0-35: mild (0-7), moderate (8-19), or severe (20-35). Estimates are based on adjusted means from the general linear model: change from baseline = baseline value and cluster and treatment.

Prostate volume was measured by transrectal ultrasound (TRUS) when biopsies were performed at Year 2 and Year 4. The investigator calculated the prostate volume using three prostate measurements (anteroposterior, cephalocaudal, and transverse diameters). Estimates are based on the adjusted means from the general linear model: log(Post-Baseline/Baseline value) = treatment and cluster and log (baseline value).

Maximum urinary flow was measured at selected sites using a Dantec Uroflow meter with a Thompson filter. Change from baseline was calculated as Month 12, 24, 36, and 48 values minus the baseline value. Estimates are based on the adjusted means from the general linear model: change from baseline = baseline Qmax and treatment. This measurement was performed at selected centers.

Medication taken during the study, including alpha blockers, was recorded at each 6-month study visit and during phone calls that occurred 3 months after each visit.

A participant was considered to have AUR when he reported being unable to urinate and required catherization. Participants were asked to report any events of AUR during the study.

A participant was considered to have a UTI if the investigator noted that the participant had UTI symptoms and had been prescribed antibiotics. Participants were asked to report any events of UTI during the study.

Participants reported events of macroscopic hematuria (visible blood in the urine) throughout the study.

Participants reported events of macroscopic hematospermia (visible blood in semen) throughout the study.

Overall survival is assessed as the number of deaths reported throughout the study.

The BII is a 4-item questionnaire that rates the level of BPH-related physical discomfort, worry, and interference with normal activities the participant has experienced. The total BII score ranges from 1 (no impact on symptoms) to 13 (major impact on symptoms). Participants completed the questionnaire at Baseline and at each 6-month visit. Participants whose language did not have a validated translation of the questionnaire did not participate. Estimates are based on the adjusted means from the general linear model:change from baseline = baseline value and cluster and treatment.

The MOS Sleep-6S is a 6-item questionnaire measuring quality of sleep. Scores range from 1 (all of the time) to 6 (none of the time) and are converted to a 1-100 scale and then averaged; a higher score indicates greater negative impact, which indicates more sleep disturbance. Participants completed the questionnaire at Baseline and at each 6-month visit. Participants whose language did not have a validated translation of the questionnaire did not participate. Estimates are based on the adjusted means from the general linear model: change from baseline=baseline value and cluster and treatment.

The NIH CSPI is a 9-item questionnaire that measures chronic prostatitis symptoms. The total score ranges from 0 to 43. A higher score indicates greater negative impact of prostatitis. Participants completed the questionnaire at Baseline and at each 6-month visit. Participants whose language did not have a validated translation of the questionnaire did not participate. Estimates are based on the adjusted means from the general linear model: change from Baseline = Baseline Value and Cluster and Treatment.

The QOL Q8 is the last question of the IPSS Questionnaire. It is a question about the participant's quality of life as it relates to prostate symptoms. Responses range from 0 (most positive) to 6 (most negative). A higher score indicates worse quality of life. Participants completed the questionnaire at Screening, Baseline, and at each 6-month visit. Estimates are based on the adjusted means from the general linear model: change from baseline = baseline value and cluster and treatment.

The PASSFI is a 3-item questionnaire that measures sexual function. Responses range from 0 (big problem) to 4 (no problem), with a total score of 12. A higher score indicates fewer problems with sexual functioning. Participants completed the questionnaire at Baseline and then yearly . Participants whose language did not have a validated translation of the questionnaire did not participate. Estimates are based on the adjusted means from the general linear model: change from baseline = baseline value and cluster and treatment.

Number of participants whose DHT, the active form of the male sex hormone testosterone, was less than 0.555 nanomoles/liter and below the level of detection at Month 48 was measured. It was measured by taking blood samples at screening and yearly thereafter.

Testosterone, a male sex hormone, was measured by taking blood samples at screening and yearly thereafter.