Abstract

OBJECTIVES:

To determine the effect of finasteride relative to placebo on prostate cancer (PCa) risk at each individual Gleason score in the Prostate Cancer Prevention Trial using a post hoc generalization of a prespecified, exploratory, biopsy sampling density-adjusted analysis.

METHODS:

The Prostate Cancer Prevention Trial enrolled 18 882 men aged >or=55 years with a prostate-specific antigen level of <3.0 ng/mL and normal digital rectal examination findings, and randomized them to finasteride 5 mg daily or placebo. PCa data from evaluable biopsies obtained within 7 years plus <or=90 days of randomization were examined. Polytomous logistic regression analysis of PCa risk was performed across individual Gleason scores using no PCa as the reference group, with no adjustment for multiplicity. The analysis model included treatment, age, race, first-degree family history of PCa, baseline prostate-specific antigen level, and the postrandomization variables of prostate volume and the number of biopsy cores at biopsy as covariates.

RESULTS:

Finasteride significantly reduced the PCa risk relative to placebo across multiple Gleason scores (4 through 7), including a 58% reduction in Gleason score 5 PCa risk (P < .0001), a 52% reduction in Gleason score 6 PCa risk (P < .0001), and a 22% reduction in Gleason score 7 PCa risk (P = .0368). Finasteride had no significant effect on the risk of Gleason score 2, 3, or 8-10 cancer.

CONCLUSIONS:

After adjusting for biopsy sampling density, finasteride significantly reduced the PCa risk relative to placebo across multiple Gleason scores in the Prostate Cancer Prevention Trial, including the most frequently detected intermediate- and high-grade cancers (Gleason scores 6 and 7).