Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Finasteride to prevent prostate cancer?

Study
Results
Prostate cancer
Other clinical events
Adverse events
Comment

It was John Maynard Keynes who commented that "in the long run we're all dead". He also said that "I would rather be vaguely right than precisely wrong". We might apply both of these to prostate cancer and its prevention. We know that it is the commonest of cancers in men, but that far more men have it than die from it, because of the variable biology and course of the disease. It is a disease of older men, and while we have various treatments for it, that begs an interesting question of whether we can prevent it.

Given that there is some evidence that androgens influence the development of prostate cancer, one strategy might be to use 5 alpha-reductase inhibitors to inhibit conversion of testosterone to the more potent androgen dihydrotestosterone to examine whether reduced androgen levels in the prostate would reduce the risk of prostate cancer. A systematic review of finasteride in BPH in trials up to four years found no evidence in support of reduced prostate cancer rates [1]. A large two-year randomised trial of the newer agent dutasteride [2] did find a significant reduction (from 1.9% to 1.1%), though with only 66 cases.

How might an investigation of prostate cancer prevention be done? It would need to be large, and of a long duration. It would need to have a placebo arm, and have good working definitions of prostate cancer both during the study, and at the end of the study. It would need to record other clinical and adverse events. A randomised trial doing just that [3] has recently reported that finasteride does reduce prostate cancer, but not without some cost in adverse events.

Study

The study enrolled men aged 55 years or older with a normal digital rectal examination, and American Urological Association score of 20 or lower (less than severe symptoms of benign prostatic hyperplasia), and a PSA of 3.0 μg/L or lower. Men were given a three-month supply of placebo tablets, and if PSA was confirmed as being 3 μg/L or lower and adherence was with 80% of nominal, they were randomised to finasteride 5 mg/day or placebo.

The planned duration was seven years, with twice yearly visits for supply of medicine, pill counts and recording of clinically significant events and adverse events. Because finasteride affects PSA levels, an end of study biopsy was planned for men without a diagnosis of prostate cancer. The size of the study was planned with loss to follow up and refusal of end of study biopsy assumptions built in.

The primary outcome was men with diagnosis of prostate cancer or who underwent end of study prostate biopsy as the intention to treat population. All PSA levels of men on finasteride were corrected by multiplying by 2.3 to account for the effects of finasteride of PSA levels. All PSA measurements and biopsy histology analyses were performed in a central laboratory.

Results

The number of men randomised was 18,880, most during the first year. The study was terminated early because of the statistical significance of the results at interim analysis, but 86% of men had completed seven years of treatment at the time of the analysis.

Prostate cancer

In the intention to treat cohort of 9,060 men, prostate cancer was diagnosed during the study or at end of study biopsy in 18% of men on finasteride and 24% of men on placebo (Table 1). The number needed to treat with finasteride 5 mg a day for seven years to prevent a diagnosis of prostate cancer in one of them was 17 (95% CI 13 to 23).


Table 1: Outcomes of prostate cancer diagnosis, other clinical outcomes, and adverse events with seven years of treatment with finasteride 5 mg daily


Percent with
(95% CI)
Outcome
Men
Finasteride
Placebo
Relative benefit
NNT
Cancer
Cancer diagnosed at any time
9060
18
24
0.8 (0.7 to 0.9)
17 (13 to 23)
Cancer diagnosed for cause during or at end of trial
3573
27
30
0.9 (0.8 to 1.0)
34 (17 to 2700)
Cancer diagnosed at end of study biopsy
7472
10
15
0.7 (0.6 to 0.8)
20 (15 to 29)
Other clinical outcomes
Benign prostatic hyperplasia
18880
5.2
8.7
0.6 (0.5 to 0.7)
28 (24 to 36)
Increased urinary urge or frequency
18880
13
16
0.8 (0.7 to 0.9)
41 (29 to 71)
Urinary retention
18880
4.2
6.3
0.7 (0.6 to 0.8)
48 (37 to 69)
Prostatitis
18880
4.4
6.1
0.7 (0.6 to 0.8)
60 (44 to 98)
Transurethral resection of prostate
18880
1.0
1.9
0.5 (0.4 to 0.7)
113 (82 to 184)
Urinary tract infection
18880
1.0
1.3
0.7 (0.6 to 0.9)
390 (150 to 670)
Percent with
(95% CI)
Adverse events
Men
Finasteride
Placebo
Relative risk
NNH
Reduced ejaculate volume
18880
60
47
1.28 (1.25 to 1.31)
7.6 (6.9 to 8.6)
Erectile dysfunction
18880
67
60
1.13 (1.11 to 1.15)
13 (11 to 16)
Loss of libido
18880
65
60
1.10 (1.08 to 1.12)
17 (14 to 23)
Gynaecomastia
18880
4.5
2.8
1.6 (1.4 to 1.9)
57 (44 to 81)


Similar efficacy was seen in men with a diagnosis for cause (clinical suspiciaon of cancer) during the trial or at the end of the study and in those in whom cancer was diagnosed at the end of study biopsy (Table 1) [note that some men might be included in both of these definitions]. Similar efficacy was seen in younger and older men (Figure 1), at different levels of starting PSA, and in men with or without prostate cancer in a first-degree relative (though cancer rates were lower in men without an affected relative; Figure 2).


Figure 1: Prostate cancer diagnosis by age at start of treatment, and treatment





Figure 2: Prostate cancer diagnosis by affected relative, and treatment




With finasteride, cancers diagnosed were more likely to be of higher grade, with Gleason scores of 7-10 (Figure 3). Most prostate cancers were clinically localised, in 98% of cases in both groups. In men who had an end of study biopsy that was not for cause and who had both a tumour and a PSA of 2.5 μg/L or less, 15% of the tumours were of high grade.


Figure 3: Gleason score 7-10 prostate cancer diagnosis




Other clinical events

Other prostate and urinary clinical events occurred less frequently with finasteride than with placebo (Table 1). The largest effect was the development of benign prostatic hyperplasia, with a number needed to treat of 28 (24 to 36) to prevent one case. Urinary retention (NNT 48) and operations for prostate resection (NNT 113) were also significantly less common with finasteride than with placebo. Mean prostate volume was 26 cc in the finasteride group compared with 34 cc in the placebo group.

Adverse events

Adverse events relating to sexual function occurred more frequently with finasteride than with placebo (Table 1). The largest difference was for reduced ejaculate volume with a number needed to harm of 7.6 (6.9 to 8.6). Erectile dysfunction and loss of libido affected 5-7% more men with finasteride than placebo, and gynaecomastia was also more frequent.

Comment

The lifetime risk of prostate cancer is about 1 in 6, which makes preventive measures attractive. Healthy living is an obvious first, and therapeutic prevention a second, choice. Treatment with finasteride 5 mg a day for seven years prevented prostate cancer diagnosis in one man for every 17 treated. Is finasteride preventing prostate cancer, or treating it? The effects were seen early, and there is at least a hint that finasteride may be treating subclinical microscopic cancer as well as delaying the onset of prostate cancer. That might accord with the finding of higher grades of cancer with finasteride, but there are many possibilities and opinions that these data do not address.

There were additional benefits with finasteride. Shrinking the prostate brought benefits like reducing the diagnosis of benign prostatic hyperplasia, and urinary retention and need for surgery. But there were also harms, notably related to the known adverse effects of finasteride on sexual function.

So choosing this as a strategy for practice is a bit early, perhaps, though it is certainly a major first step in thinking about cancer prevention. And then there's the little matter of cost. We would need to treat 17 men with finasteride for seven years to prevent a prostate cancer diagnosis in one of them. The drug costs at 2003 prices would be about £38,500 to achieve that. But finasteride will come off patent in the next few years, and lower generic prices could reduce that substantially, unless that is a vain hope.

What about looking at it from the point of view of an individual man, weighing up the benefits and risks? At 55 years old, and with no affected relatives, the chance of diagnosis of prostate cancer from the overall trial results is about 20 in 100 over the next seven years. But only about 40% of the diagnoses were made for cause during the seven years. Some of the cancers might never proceed to clinical diagnosis. That reduces the individual risk to 8 chances in 100. Balanced against possible earlier loss of sexual potency, treatment might not be a good choice. For an older man with an affected relative, a moderate prostate symptom score and less interest in sexual potency, treatment might make sense. It's all a question of how vaguely we are right.

References:

  1. JE Edwards, RA Moore. Finasteride in the treatment of clinical benign prostatic hyperplasia: a systematic review of randomised trials. BMC Urology 2002 2: 14 (http://www.biomedcentral.com/1471-2490/2/14 )
  2. CG Roehrborn et al. Efficacy and safety of a dual inhibitor of 5-alpha-reductase types 1 and 2 (dutasteride) in men with benign prostatic hyperplasia. Urology 2002 60: 434-441.
  3. IM Thompson et al. The influence of finasteride on the development of prostate cancer. New England Journal of Medicine 2003 349: 213-222.

previous story