Finasteride Does Not Increase Risk for High-Grade Prostate Cancer

Roxanne Nelson

July 13, 2009

July 13, 2009 — Data from a new trial might help quell the fear that treatment with finasteride (Proscar, Propecia) increases the risk for high-grade prostate cancer.

As previously reported by Medscape Oncology, results of the Prostate Cancer Prevention Trial (PCPT) showed that finasteride treatment reduced the prevalence of prostate cancer by 25% when used as a chemopreventive agent in patients at high risk for the disease. At the same time, it also appeared to increase the incidence of high-grade prostate cancer.

The results of a new study, published online July 7 in Clinical Cancer Research, may shed some light on the controversial findings of the PCPT. They suggest that finasteride did not cause high-grade prostate cancers but simply made them easier to diagnose.

This should make finasteride more attractive as a chemopreventive drug.

"Our results suggest that the PCPT findings, in which finasteride use was associated with an increase in diagnoses of high-grade cancers, was likely the result of a detection bias rather than an increase in de novo high-grade prostate cancer," said first author Christopher Elliott, MD, PhD, a urology resident at Stanford University, in California. "These results would suggest that the fear of promoting high-grade prostate cancer with finasteride is likely unfounded. This should make finasteride more attractive as a chemopreventive drug, especially given that finasteride is currently the only agent ever shown in a randomized trial to have a potential chemopreventive effect in inhibiting prostate cancer."

The researchers found that the level of prostate-specific antigen (PSA) declined steadily as prostate volume increased. At the same time, the level of PSA increased as prostate volume decreased. The results suggested that PSA performs better in smaller prostates for detecting both low- and high-grade disease, Dr. Elliott told Medscape Oncology.

They "also suggested that the chance of finding high-grade cancer was increased in men with smaller prostate glands, compared to those with larger glands," he said. "Finasteride has been shown to reduce prostate size by approximately 25% in men using the drug. Thus, our findings suggested that finasteride, by reducing prostate size, increased the likelihood of finding high-grade cancer in these men."

Other causes of an increased rate of high-grade cancer in men taking finasteride might exist, he added, but these were not supported by their data.

Inconsistent Findings in the PCPT

The PCPT tracked nearly 19,000 men over a 7-year period who were randomized to either finasteride or placebo. Although there was a 25% decrease in prostate cancer incidence among men taking finasteride, there was also a small but statistically significant increase in the rate of high-grade disease, compared with those taking placebo.

However, the increase in high-grade disease wasn't consistent across all groups and appeared to be limited to patients who underwent "for-cause biopsies," the researchers note. A biopsy was obtained during the PCPT for 1 of 2 reasons: "for cause," which patients underwent if they had an abnormal digital rectal exam or had a corrected PSA of 4 ng/mL or greater; or "end of study," which included all men not diagnosed with cancer during the trial. The latter group did not have abnormal digital rectal exams or elevated PSA levels.

There was a sharp difference between the 2 groups. Virtually no difference was seen in the diagnosis of high-grade cancers in the end-of-study biopsy cohort, which was 2.5% in the finasteride group and 2.3% in the placebo group. But the authors observed that there was a large difference in the for-cause biopsy group (11.5% in the finasteride group vs 7.7% in the placebo group). Thus, they write, the for-cause biopsy group drove the observation of an increased prevalence of high-grade cancer in the finasteride group.

Finasteride has been shown to decrease prostate volume by approximately 25%. In the PCPT, men in the finasteride group had a median prostate size of 25.5 cc at the time of biopsy, whereas men in the placebo group had a median prostate size of 33.6 cc. To better understand how volume changes in PSA test performance characteristics contribute to the PCPT findings, Dr. Elliott and colleagues conducted a retrospective review of 1304 men who had been referred for initial biopsy at their institution.

All of the patients had a PSA between 4 and 10 ng/mL or an abnormal digital rectal examination, and none of the participants had undergone a previous prostate biopsy or surgical therapy for benign prostatic hypertrophy. In addition, none of the patients used testosterone or 5-alpha reductase inhibitors.

Diagnosis of Cancer Improved as Prostate Volume Declined

To best approximate the PCPT study population, all of the men in this cohort, the authors note, had a PSA level of 10 ng/mL or less. Receiver-operator curves and positive predictive values (PPV) were ascertained for PSA stratified by diagnosis and prostate volume.

Within this cohort, 497 patients (38.1%) were diagnosed with prostate cancer; 247 (19%) were high grade (Gleason score ≥7) and 84 (6.4%) had a Gleason score of 8 or above. They noted that as prostate volume increased, the rates of any cancer diagnosis decreased in a statistically significant fashion.

The performance for the diagnosis of any prostate cancer improved as prostate volume decreased (area under the curve [AUC] <30 cc = 0.758; AUC 30–50 cc = 0.629; AUC >50 cc = 0.520), and this trend was statistically significant. A similar trend was seen for high-grade disease (AUC <30 cc = 0.712; AUC 30–50 cc = 0.639; AUC >50 cc = 0.497). Although it didn't reach statistical significance for the comparison of the <30 cc group to the 30–50 cc group (= .123), it did become statistically significant for the comparison of the 30–50 cc group to the >50 cc group (P =.008).

The authors also found that the PPV of a PSA greater than 4 ng/mL was also affected by prostate volume, and trends for Gleason scores of 6 or less decreased as prostate volume increased (PPV <30 cc = 25.0%; PPV 30–50 cc = 23.8%; PPV >50 cc = 17.3%). A more significant trend was observed for high-grade disease (PPV <30 cc = 39.0%; PPV 30–50 cc = 22.3%; PPV >50 cc = 10.7%).

The researchers concluded that in their referral-based cohort of men undergoing extended-scheme prostate needle biopsy, the ability of PSA to detect any cancer or high-grade cancer is affected by prostate size. "PSA performance is significantly better in prostates less than 30 cc in size as compared to those greater than 30 cc in size," they write. "This finding could explain the differing high-grade prostate cancer diagnosis rates in the for-cause biopsy cohort of the PCPT, and supports the possibility of detection bias within the trial."

Finasteride Should Be Used if Deemed Beneficial

After returning to their data and analyzing their own results, the original cancer trial researchers reached similar conclusions, said Catherine Tangen, DrPH, the statistical principle investigator for the aPCPT from the Fred Hutchinson Cancer Research Center in Seattle, Washington, in a statement. She warned that without removing or analyzing the prostates of patients in the current trial, the true prevalence of undetected prostate cancer remains unknown and leaves the actual sensitivity of the PSA test open to question.

However, she noted that these observations "are consistent with everything we found," and suggest that men "should be given the opportunity" to take finasteride if they and their doctors agree that it may be beneficial to them.

The researchers have disclosed no relevant financial relationships.

Clin Cancer Res. 2009;15:4694-4699.

Comments

3090D553-9492-4563-8681-AD288FA52ACE
Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as:

processing....