publication date: Feb. 1, 2019
20-year follow-up data are in: Prostate cancer prevention works; concerns about high-grade disease dismissed
Ian Thompson, Jr.
Principal investigator, Prostate Cancer Prevention Trial, SWOG Cancer Research Network
President, CHRISTUS Santa Rosa Hospital – Medical Center
Managing biostatistician, Statistics and Data Management Center, SWOG Cancer Research Network
Cancer Prevention Program, Fred Hutchinson Cancer Research Center
Deputy director, Statistics and Data Management Center, SWOG Cancer Research Network
Member, Cancer Prevention Program, Fred Hutchinson Cancer Research Center
Prostate cancer is the most common solid tumor in men. It has been estimated that 60-75 percent of men will have histologic evidence of prostate cancer during their lifetime and that 2-4 percent of men will die of the disease. African American men are at a greater risk of diagnosis and death.
While early detection has been found to reduce death from the disease, it comes at a great human cost: 781 men must be screened and 27 men must be treated to prevent one prostate cancer death.
Treatment is associated with significant morbidity, including impotence and incontinence. In the absence of early detection, most men who present with prostate cancer have metastatic disease; once metastases develop, most men will die of prostate cancer within 10 years.
The potential of prostate cancer prevention became a possibility in the early 1990’s. This concept occurred due to a confluence of events. With the advent of prostate specific antigen (PSA) testing in the mid to late 1980s, the rate of prostate cancer more than doubled in the U.S.
Concurrently, in 1992, the drug finasteride (Proscar) was demonstrated to effectively treat symptoms of prostate enlargement (BPH) and was ultimately found to reduce the risk of complications of BPH. Finasteride is an inhibitor of the enzyme five alpha-reductase that recapitulates a genetic mutation associated with an absence of development of BPH or prostate cancer.
Another clinical effect of finasteride was the prevention of development of male pattern baldness. With its potential effect on cancer prevention, the Board of Scientific Counsellors of the Division of Cancer Prevention (DCP) of the National Cancer Institute (NCI) recommended the exploration of the first NCI-sponsored clinical trial to determine if prostate cancer could be prevented.
DCP leadership invited Dr. Charles Coltman, group chair of the Southwest Oncology Group (now SWOG Cancer Research Network) to Bethesda to explore possible design options. Dr. Ian Thompson joined Dr. Coltman as the urologic oncologist from SWOG along with Dr. Polly Feigl and Dr. Brent Blumenstein from the SWOG Statistics and Data Management Center.
Others in attendance during the first meeting at the Executive Plaza building included Dr. Otis Brawley (former chief medical officer, American Cancer Society), Dr. Leslie Ford, Dr. Barnett Kramer, and Dr. Peter Greenwald, all from DCP. (Notably, Dr. Coltman passed away in late 2018.) (The Cancer Letter, Dec. 7, 2018)
During this meeting on May 13, 1992, multiple study designs were considered. The initial consideration of prostate cancer mortality as an endpoint was deemed unfeasible as up to 100,000 subjects and 25-30 years of follow-up would be required for such a study.
Since a prostate cancer diagnosis at that time was associated with considerable morbidity (treatment and the side effects of treatment), the study endpoint of prostate cancer prevalence was selected as one that was meaningful and achievable.
A complicating factor that needed to be accounted for in the design was that the interventional agent (finasteride) affected the primary method of prostate cancer detection (elevated PSA levels), decreasing PSA by about 50 percent and shrinking the prostate gland effecting the sensitivity digital rectal exam (DRE).
The final study design included adjusted blinded PSA values, but because the precise adjustment of PSA was not possible and the DRE effect could not be accounted for, it was determined that the only method of minimizing detection bias caused by the finasteride was to include an end-of-study prostate biopsy in all subjects at the end of their seven-year course of treatment with study drug.
After approval of the study design by the NCI and identification of more than 200 study sites around the U.S. and one site in Canada, training of hundreds of study personnel and principal investigators was conducted in 1993 followed by a press conference at the National Press Club in Washington, D.C.
Within hours, the Cancer Information Service telephone lines were flooded with requests for more information from men interested in study participation.
The PCPT was designed to accrue 18,000 men over a three-year period; due to this intense interest, not only was subject accrual complete in the anticipated three-year period but ultimately, 18,882 men enrolled and were randomized between January 1994 and May 1997.
Over the course of the study, an independent Data Safety and Monitoring Committee evaluated the critical assumptions of the study, adherence of study subjects to the trial, and, importantly, rates of recommended biopsies and prostate cancer in the two study groups: placebo and finasteride.
Two-thirds of the study subjects were randomized in the first year, and seven years later, in 2001, two-thirds of the study subjects were scheduled for their end-of-study biopsy.
On Feb. 21, 2003, during a DSMC meeting, on the basis of an analysis that concluded that further biopsies would not change the results that were reached at that time, the DSMC voted to recommend study closure and to discontinue any further treatment for study subjects who were still on study.
The conclusion was that finasteride significantly reduced the relative risk of prostate cancer, by 24.8 percent, meeting the primary objective of the study. A total of 1,147 prostate cancers were seen with placebo, compared with 803 with finasteride.
Complicating the study finding was a paradox: while prostate cancer risk was reduced significantly, a greater number of high-grade prostate cancers were seen with finasteride. While there were 344 fewer overall cancers, there were 43 more high-grade cancers with finasteride.
The study leadership team decided that the best way to report these findings and to close the study would be to prepare a manuscript for publication that would then go through high-level peer review to optimize the scientific discussion of this paradoxical finding and to release information to participants and institutions concurrently with a fast-track publication.
On July 17, 2003, The Influence of Finasteride on the Development of Prostate Cancer was published in New England Journal of Medicine (NEJM). The publication related the significant reduction in risk of prostate cancer, the smaller but statistically-significant increased risk of high-grade cancer, but did not address analyses to explore why these findings may have occurred; these analyses would have to wait for additional data to be collected, cleaned and analyzed.
Concurrent with the NEJM manuscript was the publication of an editorial by Dr. Peter Scardino. In that editorial, Dr. Scardino stated that finasteride should not be recommended for prevention of prostate cancer and that further follow-up would be required to understand the study’s findings.
Side effects of finasteride included a small but statistically significant negative impact on sexual function and a small increased risk of breast enlargement.
Over the ensuing years, three key investigations and publications helped us understand why there were more high-grade cancers seen in the finasteride study group.
The first of these was in 2006, when PCPT investigators reported in the Journal of the National Cancer Institute (JNCI) that finasteride significantly improved the performance of PSA for prostate cancer detection and for detection of high-grade prostate cancer.
In 2007, the PCPT team reported in Journal of Urology that finasteride similarly significantly improved the performance of digital rectal examination for detection of prostate cancer.
Finally, in 2007, a group led by Dr. Scott Lucia, the lead PCPT study pathologist, reported in JNCI on the comparison of biopsy and radical prostatectomy tumor grades in the two study arms.
They found that if high grade prostate cancer were truly present (confirmed by radical prostatectomy), it was missed on biopsy 50 percent of the time in men who received placebo, compared to only 30 percent of the time in men who received finasteride.
Several subsequent studies evaluated how these biases (that increased the likelihood of detecting prostate cancer and detecting high-grade cancer in men who received finasteride) would have affected the final study results.
In 2008, in the journal Cancer Prevention Research, SWOG biostatistician Dr. Mary Redman estimated that accounting for these biases, the overall risk of cancer was reduced by 30 percent, and that, although based on small numbers, the risk of high-grade cancer was reduced by 28 percent with finasteride.
In 2010, a group of investigators working with Glaxo Smith Kline, published in New England Journal of Medici the results of the REDUCE (Reduction by Dutasteride of Prostate Cancer Events) clinical trial. This study examined the impact of a dual 5-alpha reductase inhibitor–dutasteride–on the risk of prostate cancer in men with elevated PSA levels (2.5 to 10 ng/mL) and a prior negative biopsy.
The authors found that dutasteride reduced the relative risk of prostate cancer by 22.8 percent over four years. While the total number of Gleason 7-10 tumors was similar in the two study arms, there were more Gleason 8-10 tumors in the men receiving dutasteride.
Based on the results of the REDUCE clinical trial, GSK approached the U.S. Food and Drug Administration about placing these results in the product information for dutasteride. Although almost 10 years previously, the significant reduction of prostate cancer that had been identified with finasteride, these results had not been placed in the product information for finasteride, initially manufactured by Merck. Notably, finasteride became a generic drug in October 2014.
In response, the FDA referred the GSK request to the Oncologic Drugs Advisory Committee (ODAC) in late 2010. Merck was requested to also present information related to finasteride at that meeting and Dr. Ian Thompson (study principal investigator) and Dr. Catherine Tangen (study lead statistician) attended.
The ultimate decision was made to not only exclude information related to a reduced risk of prostate cancer with both drugs, but to also place a warning in the product information materials for finasteride stating that there may be an increased risk of high-grade prostate cancer.
The information related to improved detection of prostate cancer and of high-grade prostate cancer was not included.
Over the years following the 2003 release of the original PCPT results that finasteride reduced the risk of prostate cancer by 24.8 percent, despite the enhanced detection of cancer with finasteride, very little interest was seen for the use of this drug to prevent the most common cancer in men.
The most common reason cited was the potential increased risk of high-grade cancer that could negate the reduced risk by increasing risk of death from prostate cancer.
In August 2013, PCPT study investigators published long-term survival outcomes of the study in NEJM, finding no differences in overall survival in men randomized to finasteride versus placebo.
Subsequently, in 2018, Dr. Joseph Unger and SWOG investigators, after linking PCPT participants with Medicare data, reported on long-term prostate cancer risk in JNCI. The investigators found that with up to 20 years of follow-up, the approximately 25 percent reduction in risk was durable.
Understanding that the primary hurdle to the use of finasteride for prevention of prostate cancer was the observed increased number of high-grade cancers and that these high grade cancers could increase the risk of prostate cancer death, SWOG investigators led by Phyllis Goodman conducted an analysis to answer the question: what was the impact of seven years of finasteride on PCPT participants risk of prostate cancer death?
To address the question, PCPT participants were linked to the National Death Index which provided cause of death.
The results were published on Jan. 24, 2019, in NEJM. With 296,842 person-years of follow-up and a median follow-up of 18.4 years, of 9,423 men randomized to finasteride, there were 3,048 deaths and 42 deaths due to prostate cancer. By comparison, of 9,457 men randomized to placebo, there were 2,979 deaths of which 56 were due to prostate cancer.
While there was a 25 percent lower risk of prostate cancer on the finasteride arm, with a small number of prostate cancer deaths, this difference was not statistically significant.
Interestingly, of the prostate cancer deaths in which tumor grade was known, more than a third of the men who died of prostate cancer were originally diagnosed with a Gleason < 6 tumors, the type of cancer that is reduced significantly with finasteride.
The majority of the results of the PCPT are now in. Seven years of treatment with finasteride reduces the risk of prostate cancer by about 25 percent and that reduced risk is durable with 20 years of follow-up. Most tumors that are prevented are low-grade, tumors that are commonly “observed” at this time, but that are ultimately treated with radical prostatectomy or radical radiation therapy in about half of men.
These treatments can have significant side effects, including impotence and incontinence.
Finasteride also improves urinary symptoms in men with BPH and significantly reduces the need for treatment of BPH or complications of BPH (such as urinary retention).
The cost of the drug is $7 to $8 per month, as it is generic.
Despite significantly fewer cancers in men who receive finasteride, the drug also improves detection of prostate cancer by better performance of PSA testing and of prostate biopsy. From the most recent analysis, we now know that overall survival of men and risk of prostate cancer death is unaffected by finasteride.
As had been anticipated, it would have taken a study many times larger to have been able to detect a statistically significant difference in risk of prostate cancer death.
Historically, one man in six has been diagnosed with prostate cancer in the U.S. A 25 percent reduction in risk with finasteride would have a profound impact on cancer risk and our nation’s war on cancer. That the drug has a relatively low risk of side effects, is inexpensive, and improves urinary function in men who commonly suffer problems from prostate enlargement makes it an even more attractive method of prevention. Men who are most likely to benefit are those who are undergoing PSA testing.
Physicians who are ordering PSA testing for men 55 years of age and older should explain these results to patients and offer the opportunity to reduce their risk of prostate cancer with finasteride.