Are we making progress in the chemoprevention of prostate cancer? The disease is an ideal target: not only is it the most common cancer among men in the United States, it has the highest age-specific incidence of any cancer. If we could prevent the disease or delay its progression, many men could be spared the burden of diagnostic procedures and treatments, and costs to the health care system could be reduced substantially. Because we lack a complete understanding of the pathophysiology of prostate cancer, chemopreventive strategies have empirically focused on two approaches: treatment with antioxidants to reduce damage to DNA and suppression of androgenic stimulation of the prostate. The latter is the focus of the article by Andriole and colleagues in this issue of the Journal.

Testosterone, the major circulating androgen in men, is converted to the major intracellular androgen, dihydrotestosterone, by steroid 5-reductase isoenzymes, designated as type 1 and type 2. The effect of finasteride, an inhibitor of the type 2 isoenzyme, on the development of prostate cancer has been reported previously.Andriole and colleagues report on the effect of dutasteride, an inhibitor of both isoenzymes, in reducing the risk of prostate cancer. This 4-year, international, placebo-controlled, randomized trial evaluated the effect of dutasteride in reducing the presence of cancer, as assessed on random biopsies performed after 2 and 4 years of treatment. They report an absolute reduction of 5.1 percentage points, and a 23% relative reduction, in incident prostate cancer detected on biopsy over the course of 4 years (a rate of 25.1% in the placebo group vs. 19.9% in the dutasteride group). As the authors point out, this finding most likely represents shrinkage or inhibition of the growth of existing tumors rather than prevention of cancer. Men who were treated with dutasteride had a significant improvement in urinary function, which was offset by a modest decline in sexual performance. This important article raises three critical questions: Is the finding of a decrease in cancer detected on random biopsies a meaningful end point? Was there a reduction in the incidence of tumors with lethal potential? Should these findings influence the way we treat our patients?

Because prostate cancer is found at autopsy in approximately half of older men who die without knowing they had the cancer, random biopsies performed without a clinical indication are not informative. Instead, in clinical practice biopsies are performed only "for cause" — that is, when there is an elevation in the serum prostate-specific antigen (PSA) level or an abnormal digital rectal examination. However, evaluating the effect of 5-reductase inhibitors on for-cause biopsies is complicated. Because these medications block hormonal stimulation of the prostate, they dramatically lower PSA levels. Thus, with the use of standard PSA cutoff points it is difficult to ascertain whether these drugs actually prevent prostate cancer or simply reduce the likelihood that men will undergo biopsies. The only study that definitely addressed this issue was the Prostate Cancer Prevention Trial (PCPT), which compared treatment for 7 years with finasteride, at a dose of 5 mg per day, with placebo. During the 7 years of the study (but not including data from the end-of-study biopsies), there was a 25% reduction in the number of cancers diagnosed for cause in the finasteride group. However, as compared with men taking placebo, a disproportionately higher percentage of men taking finasteride for a clinical indication (an abnormal examination or PSA level corrected for the effect of finasteride) did not undergo the recommended biopsy. Of the men who actually underwent a biopsy, cancer was detected in 26.5% of those taking finasteride and 29.5% receiving placebo, a 10% reduction that was not statistically significant (relative risk, 0.90; 95% confidence interval, 0.81 to 1.00). Furthermore, by the end of the study at year 7, there was no between-group difference in the number of cancers. Supporting this observation, a recent observational study from the Finnish Prostate Cancer Trial (Current Controlled Trials number, ISRCTN49127736 showed that there was no significant decrease in the incidence of prostate cancer among men in the screening group who were taking finasteride, as compared with the men who were receiving placebo. In the study by Andriole and colleagues, of the biopsies performed for cause outside the protocol, 16.6% in the dutasteride group and 16.7% in the placebo group were positive. Thus, neither dutasteride nor finasteride significantly reduced the risk of prostate cancer among men who were followed closely and who underwent a biopsy because of an elevated PSA level (corrected for the effect of the drug) or an abnormal digital rectal examination; unfortunately, this is the setting that would be used for prevention.

Were the cancers that were prevented clinically significant? The percent of positive biopsies in the placebo group (25%) substantially exceeded the published lifetime risk of prostate cancer (17%), suggesting that many of the tumors were likely to have been clinically insignificant. Furthermore in the dutasteride group, the reduction in the rate of incident cancer was limited to the incidence of prostate tumors with Gleason scores of 5 to 6, which are moderately well differentiated; there was no significant reduction in the incidence of tumors that were less differentiated (Gleason scores of 7 to 10) — tumors that are more likely to be lethal. The finding that dutasteride was ineffective in reducing these high-grade tumors is somewhat disappointing, particularly in light of some previous studies suggesting that there may be increased levels of 5-reductase type 1 in these tumors.Similarly, in the random biopsies performed at the end of the PCPT, there was no decrease in high-grade disease. On the basis of the hypothesis that it may be easier to find high-grade cancer on biopsy when prostate volumes are reduced, post hoc mathematical models suggested that there may have been a reduction in the incidence of tumors with Gleason scores of 7 to 10 in the PCPT among men in the finasteride group; nevertheless, the results of Andriole and colleagues were unchanged when these volume-related adjustments were applied.

What are the major therapeutic effects of 5-reductase inhibitors? Among men without cancer, both finasteride and dutasteride treat lower urinary tract symptoms, produce dramatic prostate shrinkage, and reduce serum levels of PSA by 50% or more. However, for the patient who believes that he is taking a drug to prevent prostate cancer, this decline in PSA level can lead to a false sense of security. Data from the PCPT indicate that finasteride produces a progressive suppression of PSA for the duration of treatment. To estimate what the "true" PSA level would be if finasteride were not taken, it is necessary to multiply the PSA level by 2.0 for the first 2 years, by 2.3 for years 2 through 7, and thereafter by 2.5. However, if the PSA level ever begins to rise at all, a biopsy should be performed, because with an increase in PSA level, the risk of cancer is increased by a factor of 3 and the risk of high-grade disease is increased by a factor of 6. In the Finnish study cited above, men who received finasteride for more than 4 years had a risk of high-grade disease that was increased by a factor of 2.6.

What advice should we be giving our patients? Dutasteride and finasteride do not prevent prostate cancer but merely temporarily shrink tumors that have a low potential for being lethal, and they do not reduce the risk of a positive biopsy in patients who have an elevated PSA level (corrected for the effect of the drug) or an abnormal digital rectal examination. Furthermore, the use of these drugs for prevention may be somewhat risky. Because PSA levels are suppressed, men may have a false sense of security, and if prostate cancer ever develops, the diagnosis may be delayed until they have high-grade disease that may be difficult to cure.