The effect of finasteride on the risk of acute urinary retention and the need for surgical treatment among men with benign prostatic hyperplasia
CommentThis was a 4-year randomized study of finasteride vs. placebo in men with moderate to severe symptoms of obstructive uropathy and clinically enlarged prostate glands. It demonstrated a significant improvement in symptoms, as well as a significant decrease in the need for surgery (5% vs. 10%) and the incidence of acute urinary retention (3% vs. 7%) over 4 years in the finasteride group. This translates into a Number Needed to Treat of 20 patients treated for 4 years to prevent one prostate operation.
Only men with enlarged prostates on rectal examination and moderate to severe symptoms were included in this study. This is important to note, since previous studies of finasteride vs. alpha-blockers for symptoms of obstructive uropathy indicate that finasteride's effect may be limited to men with enlarged prostate glands. A study published in the NEJM in 1996, which did not require evidence of prostatic enlargement on physical exam, found no benefit for finasteride, but a significant benefit of terazosin. In an editorial published along with that study, Walsh hypothesizes that obstructive symptoms may be due to smooth muscle hyperplasia or to glandular hyperplasia. Men with smaller prostates and obstructive symptoms are more likely to suffer from smooth muscle hyperplasia and would not be expected to benefit from finasteride but would benefit from an alpha blocker. Men with large prostates are more likely to have glandular hyperplasia, which would be expected to respond to an anti-testosterone agent such as finasteride. Thus, the results presented here cannot be automatically extrapolated to all men with obstructive uropathy symptoms, only to those with clinically enlarged prostates.
One point not addressed in the article is the serum PSA level. In this study, men with mildly elevated PSA levels were required to undergo a prostate biopsy in order to participate in the study. This probably caused a bias towards lower PSA levels among study patients, since many patients with only mildly elevated PSA levels would be reluctant to undergo this procedure. In fact, despite the fact that the average prostate volume was elevated (55 ml), the average serum PSA was only 2.8 ng/ml. This could conceivably have biased the results, possibly against finasteride, since an elevated PSA level could be a marker for glandular tissue responsive to the drug. It would have been instructive to know the effect of serum PSA levels on the study results.
It is interesting that the primary endpoint of this study was the improvement
in symptoms with finasteride, whereas the most important results were represented
by the secondary endpoints (surgery and urinary retention). At the time
the study was initiated, it was already known that finasteride significantly
improved symptoms in men with enlarged prostate glands. Thus, the choice
of primary endpoint was presumably made to ensure a positive result of
the trial, while leaving open the possibility of emphasizing the importance
of the secondary endpoints if they turned out to be significant, as was
the case. Although this is not a major problem, the primary endpoints of
a study are those upon which sample size determinations are usually made,
and should represent the main topic being studied.
March 9, 1998
ReferencesReferences related to this article from the NLM's PubMed database.
Reader CommentsDate: Fri, 20 Mar 98
The article posted was very interesting. I have even found finasteride produced some symptomatic improvement in what turned out to be a definite carcinoma. I don't think its use should mask the diagnosis as I usually, as a general practitioner, take a PSA before starting treatment. I also usually refer to a urologist for biopsy etc. In my experience finasteride is well tolerated.
John Warre. (UK)
Thu, 26 Mar 1998
I have two comments:
How sensitive is digital prostatic examination (DPE) to identify an enlarged prostate gland? Are there any studies available? If not, then the decision to include patients in the study should have been based on objectively (sonogram/MRI) documented gland size enlargement instead of DPE i.e. any patient with symptoms of prostatism should have been subjected to these exams before inclusion in the study. Because we often encounter patients with normal prostate size (muscular component predominence hypothetically) who may have an enlarged prostate missed by DPE. That would have helped in assesing sensitivity and specificity of DPE as well, considering Sonogram/MRI as gold standard.
How long does it take for finasteride to become effective once we start
it on a patient with Benign Prostatic Hypertrophy (BPH) ( given the hypothesis
that it decreases glandular component of Prostate)? What should be done
for patients with symptoms of prostatism meanwhile? Use alphablockers?
We may have to reconsider our options. Is finasteride overrated for BPH?
I am glad it has found another medical use [for baldness. mj].
Alpha blockers are likely to be effective more rapidly than finasteride, and may need to be given initially along with finasteride if rapid symptomatic improvement is desired. --mj
Mon, 22 Feb 1999
Evidence based medicine uses the term number needed to treat (NNT) very
often to discuss the clinical significance of a therapy. It is defined
as the number of patients needed to be treated in order to obtain one benefit.
NNT is calculated as 100 / (absolute risk reduction). In the Study by McConnell
et. al the absolute risk reduction due to the use of finasteride was 7%-4%=
3%, and the NNT is 100/3=33. Despite the statistical significance, one
needs to treat 33 patients with finasteride to get 1 fewer episode of urinary
retention. The NNT for prostate surgery was 20, which seemed to be
clinically more relevant information. The NNH (number needed to harm) was
about 23 for impotence during the first year of treatment. One needs to
use these numbers when discussing the risks versus benefit of finasteride
therapy with patients. Both groups were equally affected by prostate cancer.
Amit Ghosh, MD
Journal Club homepage