February 1, 2008 — Blood levels of androgens and other sex hormones do not seem to be related to the risk for prostate cancer. The finding comes from a huge pooled analysis of data from 18 studies, published online in the January 29 Advance Access issue and will appear in the February 6 issue of the Journal of the National Cancer Institute. It "confirms the lack of evidence to support an androgen–prostate cancer hypothesis," according to an accompanying editorial.

A link between elevated androgen levels and an increase in the risk for prostate cancer has been widely hypothesized, despite limited supportive epidemiologic evidence, comment Paul Godley, MD, PhD, and colleagues from the University of North Carolina at Chapel Hill, in an accompanying editorial. Androgens play a key role in the development of the prostate, and androgen suppression in advanced cancer can lead to a dramatic regression of the disease, they point out. Hence, there has been an "intense and sustained interest in confirming an androgen-driven hypothesis," they add.

The new analysis provides no evidence of such a link, however. An international collaboration of researchers, headed by Andrew W. Roddam, DPhil, from the University of Oxford, Oxford, United Kingdom, took another look at the original data collected in the 18 studies, consisting of 3886 men with prostate cancer and 6438 controls. Each of the studies had looked at the relationship between androgens and the risk for prostate cancer, but the results were inconclusive; some suggested a positive association, but many of the studies had limited power, the researchers comment. In the new analysis, all of the data from these previous studies were pooled together. The team looked at blood samples taken before the men developed prostate cancer and analyzed serum concentrations of testosterone, free testosterone, dihydrotestosterone (DHT), dehydroepiandrosterone sulfate (DHEA-S), androstenedione, androstanediol glucuronide, estradiol, and calculated free estradiol. They found no significant relationship between the serum levels of any of these hormones and the risk of developing prostate cancer.

This is an "impressive pooled analysis" that enhances "our understanding of prostate cancer epidemiology," the editorialists comment. It also offers a new opportunity because it now "obliges the scientific community to move past a seductive, clinically relevant, and biologically plausible hypothesis and get on with the difficult task of exploring, analyzing, and characterizing modifiable risk factors for prostate cancer."

However, the question of where to go from here will take some serious discussion and debate, Dr. Godley commented to Medscape Oncology. "I'm pretty sure that there is not an obvious answer to what are the most likely causes of prostate cancer," he said.

The finding also raises questions about the use of finasteride for the prevention of prostate cancer. This drug, a 5-alpha reductase inhibitor, blocks the conversion of testosterone into the more potent DHT. In the Prostate Cancer Prevention Trial, previously reported by Medscape Oncology, finasteride reduced the occurrence of prostate cancer by approximately 25% during a 7-year period, although the risk for high-grade tumors was higher in the treated vs the untreated group. These results offer a "tantalizing possibility for effective prostate cancer chemoprevention," add the editorialists, but the hypothesis that provided the theoretical foundation for this trial has now been "convincingly debunked" by the latest finding.

However, this finding will probably not affect the use of finasteride for prostate cancer prevention, although it undermines part of the rationale for its use, Dr. Godley commented to Medscape Oncology. "I'm not sure that finasteride is being used much for that purpose anyway." In addition, the theory is a little different in each of these 2 trials, he pointed out. The current study set out to investigate whether high androgen levels increase the risk for prostate cancer and showed that they do not, whereas the finasteride trial set out to determine whether lowering androgen levels reduces the risk for prostate cancer and showed that it did.

Neither the authors nor the editorialists have disclosed any relevant financial relationships.

J Natl Cancer Inst. Published online January 29, 2008. 2008;100:158-159, 170-183.

Clinical Context

Study Highlights

• Studies were identified by literature searches of multiple databases.
• 18 studies that included 3886 men with incident prostate cancer and 6438 controls representing more than 95% of the worldwide data were identified.
• 14 of the 18 studies used a matched case–control design nested within a prospective cohort collection or a randomized trial.
• Blood samples were collected from apparently healthy men who were then observed to identify who developed prostate cancer.
• The same blood sample used for prostate-specific antigen testing was also used for determining serum hormone concentrations.
• 3 of the 18 studies were carried out as a full cohort of case-cohort analyses in which hormone assays were performed on stored serum from some or all cohort participants.
• Principal investigators provided data on case-control status, age, date of recruitment, interval between blood collection and diagnosis of prostate cancer, demographics, and risk factors.
• Free testosterone and free estradiol were calculated from the reported measured serum concentrations of testosterone or estradiol and sex hormone–binding globulin (SHBG) by use of law of mass action, assuming a constant serum albumin concentration of 43 g/L.
• Each study provided grade and stage of prostate cancer if available.
• Data on serum testosterone concentrations were available for 3886 cases with prostate cancer and 6438 controls, with a similar amount of data available for SHBG, enabling the free testosterone to be calculated for 3550 cases and 5815 controls.
• Data on estradiol were available for 2186 cases and 3039 controls.
• Androstanediol glucuronide was available for 2453 cases and 3035 controls.
• Individual data was available for DHT, DHEA-S, and androstenedione for 1000 cases.
• Among controls, mean age at recruitment was 46 to 72 years, mean body mass index was 22.4 to 28.2 kg/m², and 4% to 37% smoked with mean alcohol consumption ranging from 11.5 to 21.7 g/day.
• There was considerable variation between time of blood test and time of diagnosis of prostate cancer.
• Most cases were diagnosed in men older than 60 years, many older than 70 years.
• Among those diagnosed with prostate cancer, 60% to 70% had localized or low-grade disease, or both.
• Serum concentrations of testosterone, free testosterone, and DHT correlated positively with each other, and androgen and estradiol levels were moderately associated with one another.
• There were no statistically significant correlations between serum concentrations of any of the androgens or estrogen and the risk for prostate cancer, comparing the highest and lowest fifth (eg, for testosterone, relative risk [RR], 0.94; 95% CI, 0.82 - 1.07 and for free testosterone, RR, 1.11; 95% CI, 0.96 - 1.27).
• There was a possible inverse association between the highest and lowest fifth of DHT concentrations, but no dose-response relationship was observed.
• SHBG was statistically significantly and inversely related to prostate cancer risk, with RR of 0.86 (95% CI, 0.75 - 0.98) comparing the highest with the lowest fifth.
• There was no evidence of heterogeneity between studies for estimates of serum hormone concentration and prostate cancer risk.
• Adjustment for age, education, body mass index, marital status, smoking, and alcohol intake did not change the findings.

Pearls for Practice

• There is no association between serum concentrations of testosterone, free testosterone, DHT, androstanediol glucuronide, and androstenedione and the risk for prostate cancer and a possible inverse association with SHBG level.
• There is no association between serum estradiol level and the risk for prostate cancer in men.

CME Test