BACKGROUND

Sex steroids, particularly androgens, have been implicated in the pathogenesis of prostate cancer. Data from previous studies comparing circulating hormone levels in men with and without prostate cancer are difficult to interpret, since the studies were limited in size, hormone levels were analyzed in blood drawn after the diagnosis of cancer, nonrepresentative control subjects were used, and hormone and hormone-binding protein levels were not simultaneously adjusted.

PURPOSE

We conducted a prospective, nested case-control study to investigate whether plasma hormone and sex hormone-binding globulin (SHBG) levels in healthy men were related to the subsequent development of prostate cancer.

METHODS

Among participants in the Physicians' Health Study who provided plasma samples in 1982, we identified 222 men who developed prostate cancer by March 1992. Three hundred ninety control subjects, matched to the case patients on the bases of age, smoking status, and length of follow-up, were also identified. Immunoassays were used to measure the levels of total testosterone, dihydrotestosterone (DHT), 3 alpha-androstanediol glucuronide (AAG), estradiol, SHBG, and prolactin in the stored (at -82 degrees C) plasma samples. Correlations between individual hormone levels and between hormone levels and SHBG in the plasma of control subjects were assessed by use of Spearman correlation coefficients (r). Odds ratios (ORs) and 95% confidence intervals (CIs) specifying the prostate cancer risk associated with quartile levels of individual hormones, before and after adjustment for other hormones and SHBG, were calculated by use of conditional logistic regression modeling. Reported P values are two-sided.

RESULTS

No clear associations were found between the unadjusted levels of individual hormones or SHBG and the risk of prostate cancer. However, a strong correlation was observed between the levels of testosterone and SHBG (r = .55), and weaker correlations were detected between the levels of testosterone and the levels of both estradiol (r = .28) and DHT (r = .32) (all P < .001). When hormone and SHBG levels were adjusted simultaneously, a strong trend of increasing prostate cancer risk was observed with increasing levels of plasma testosterone (ORs by quartile = 1.00, 1.41, 1.98, and 2.60 [95% CI = 1.34-5.02]; P for trend = .004), an inverse trend in risk was seen with increasing levels of SHBG (ORs by quartile = 1.00, 0.93, 0.61, and 0.46 [95% CI = 0.24-0.89]; P for trend = .01), and a non-linear inverse association was found with increasing levels of estradiol (ORs by quartile = 1.00, 0.53, 0.40, and 0.56 [95% CI = 0.32-0.98]; P for trend = .03). No associations were detected between the levels of DHT or prolactin and prostate cancer risk; for AAG, a marker of 5 alpha-reductase activity, only suggestive evidence of a positive association was found. The results were essentially unchanged when case patients diagnosed within 4 years of plasma collection, case patients diagnosed with localized (i.e., nonaggressive) disease, or control subjects with elevated prostate serum antigen levels (> 2.5 ng/mL) were excluded from the analyses.

CONCLUSIONS

High levels of circulating testosterone and low levels of SHBG-both within normal endogenous ranges-are associated with increased risks of prostate cancer. Low levels of circulating estradiol may represent an additional risk factor. Circulating levels of DHT and AAG do not appear to be strongly related to prostate cancer risk.