Study Shows Testosterone Levels Predict Prostate Cancer Risk

BETHESDA, Md--Physicians have long hypothesized that natural variationin sex hormones may influence prostate cancer risk, said MeirStampfer, MD, of the Harvard School of Public Health. Effortsto prove this, however, have yielded unclear results. Now, byexamining the interrelationship of different sex hormones, Dr.Stampfer and his colleagues have achieved what he calls the firstclear demonstration that circulating levels of sex hormones canpredict a man's risk of developing prostate cancer.

The finding further suggests the need for studies of how lifestylepractices might change circulating hormone levels, he said atthe General Motors Cancer Research Foundation annual conference.

Previous studies failed to demonstrate a convincing link for avariety of reasons, Dr. Stampfer said. Some were simply too small.Often the blood used was collected after diagnosis of prostatecancer, and concern lingered that the cancer therapy or the canceritself had altered sex hormone levels among the participants inthe study.

Dr. Stampfer and his colleagues, who included Peter H. Gann, MD,ScD, turned to the Physicians' Health Study. In 1982, 14,916 ofthe 22,071 male US physicians in the study, all with no priorcancer, provided samples of their blood, which were frozen andstored. Study participants answered follow-up questionnaires every6 months. The Harvard team matched 222 physicians who developedprostate cancer in the following decade with 390 healthy, cancer-freeparticipants.

Blood from each man was thawed and measured for testosterone;DHT (dihy-drotestosterone, the active form of testosterone inthe prostate); AAG (3-alpha-androstanediol glucuronide, an androgenicmetabolite of testosterone); estradiol; prolactin; and SHBG (sexhormone binding globulin).

Strong Intercorrelations

"Our thinking was that the androgens--testosterone, DHT,and AAG--would be associated with an increased risk of prostatecancer and that estradiol would be associated with a decreasedrisk," he said. Since SHBG "binds more to testosteronethan to estradiol, we thought that the net effect of SHBG wouldbe to reduce prostate cancer."

Individually, the hormones revealed only a "weak or nonexistentassociation" with risk. However, circulating sex hormoneshave strong intercorrelations, Dr. Stampfer noted. For example,men with high testosterone usually have relatively high estrogenlevels as well; SHBG binds both testosterone and estrogen andthus can inhibit the biological effects of both hormones in thebody.

So the Harvard researchers looked at the levels "simultaneously,"using a mathematical model that statistically controlled for theircomplex relationships. "That permitted us to see very clearlya sharp and rather strong effect of male hormones raising riskand female hormones lowering risk," Dr. Stampfer said.

Using conditional logistic regression models, the team calculatedodds ratios for developing prostate cancer based on hormone levels.For testosterone, the risk of prostate cancer rose from 1.00 formen with levels that put them in the lowest quartile to 1.41,1.98, and 2.60 in the upper quartiles.

SHBG showed an inverse trend, from 1.00 to 0.93, 0.61, and 0.46.Estradiol produced a nonlinear inverse association: 1:00 at thelowet quartile and then 0.53, 0.40, and finally 0.56 at the topquartile.

No risk was associated with DHT or prolactin. AAG yielded only"suggestive evidence" that it increased prostate cancerrisk. "The DHT results seem somewhat surprising because weexpected that to show risk," Dr. Stampfer said.