Relationships between Circulating and Intraprostatic Sex Steroid Hormone Concentrations.

Sex hormones have been implicated in prostate carcinogenesis, yet epidemiologic studies have not provided substantiating evidence. We tested the hypothesis that circulating concentrations of sex steroid hormones reflect intraprostatic concentrations using serum and adjacent microscopically verified benign prostate tissue from prostate cancer cases. Incident localized prostate cancer cases scheduled for surgery were invited to participate. Consented participants completed surveys, and provided resected tissues and blood. Histologic assessment of the ends of fresh frozen tissue confirmed adjacent microscopically verified benign pathology. Sex steroid hormones in sera and tissues were extracted, chromatographically separated, and then quantitated by radioimmunoassays. Linear regression was used to account for variations in intraprostatic hormone concentrations by age, body mass index, race, and study site, and subsequently to assess relationships with serum hormone concentrations. Gleason score (from adjacent tumor tissue), race, and age were assessed as potential effect modifiers. Circulating sex steroid hormone concentrations had low-to-moderate correlations with, and explained small proportions of variations in, intraprostatic sex steroid hormone concentrations. Androstane-3α,17β-diol glucuronide (3α-diol G) explained the highest variance of tissue concentrations of 3α-diol G (linear regression = 0.21), followed by serum testosterone and tissue dihydrotestosterone ( = 0.10), and then serum estrone and tissue estrone ( = 0.09). There was no effect modification by Gleason score, race, or age. Circulating concentrations of sex steroid hormones are poor surrogate measures of the intraprostatic hormonal milieu. The high exposure misclassification provided by circulating sex steroid hormone concentrations for intraprostatic levels may partly explain the lack of any consistent association of circulating hormones with prostate cancer risk. .