Brawley et al. on prostate cancer screening


As mentioned on Tuesday this week, an article and an editorial on prostate cancer screening were just published in CA: A Cancer Journal for Clinicians. Initial media reports on the article and editorial implied that the authors were suggesting that there is no value to testing for prostate cancer. Let us be very clear. This is not what the article or the editorial say at all.

The article (“Screening for prostate cancer“) is by Brawley, Ankerst, and Thompson, and the full text is now available on line. The editorial (“Prostate cancer: current evidence weighs against population screening“) is by Boyle and Brawley, and the full test of this editorial is also available.

The article provides a thorough review of widely known data and makes the following key points:

  • Since the mid-1980s, the use of PSA testing has more than doubled the likelihood that a man will have a prostate cancer diagnosis in his lifetime.
  • There has been a significant decrease in the annual mortality rate from prostate cancer during the same time period, but the degree to which this decrease in mortality is strictly a consequence of the increase in numbers of men diagnosed and the timing of those diagnoses is not well established.
  • The results from the PLCO and the ERSPC screening trials, published earlier this year and widely publicized and discussed, demonstrate that, at best, the impact of widespread, population-based, prostate cancer screening on overall and prostate cancer-specific mortality is small.
  • There are significant potential adverse consequences associated with early testing for and consequent early diagnosis of prostate cancer.
  • Men who, for whatever reason, are undergoing regular testing for risk of prostate cancer, can reduce that risk if they receive treatment with finasteride (or, in fact, dutasteride).

The justifications for all of these points are well-established in the article, and The “New” Prostate Cancer Cancer InfoLink has no criticism of any of these points.

It is also worth noting some additional points that will be of no comfort whatsoever to patients with prostate cancer who were initially diagnosed when they < 50 years of age, but they are true, nonetheless:

  • About 2.5 percent of all white men diagnosed with prostate cancer are < 50 years of age at diagnosis.
  • About 5.2 percent of all black men diagnosed with prostate cancer are < 50 years of age at diagnosis.
  • For the period 2001-05, the median age at diagnosis for white Americans was 68 years.
  • For the period 2001-05, the median age at diagnosis for black Americans was 65 years.
  • For the same time period, the median age at death from prostate cancer was 80 for whites and 77 for blacks.

This means that the median time from diagnosis to death for the men who actually die of prostate cancer is now about 12 years, and that (overall, on average) prostate cancer has minimal impact on life expectancy. Again, we note that such statistics are of no comfort to those men who die young with aggressive prostate cancer or who have suffered for years from the adverse effects of their treatment(s).

What the article actually makes clear is that we just do not know how best to use the tools currently available in order to test an individual man so as to ascertain with accuracy his real risk for clinically significant prostate cancer. We may be able to determine that he has prostate cancer cells in his prostate. And we certainly can tell, if we find those cells, whether they may be aggressive (if there is Gleason grade 4 or 5 prostate cancer tissue visible on pathology slides). But we still have no good way to tell what to do with that patient just because he has these cells present in his prostate. So — surprise, surprise — we need better tests, as America’s Prostate Cancer Organizations have already clearly stated.

Last but not least:

  • The article attempts to give what little useful advice it can to physicians about how to discuss the appropriateness of prostate cancer screening with individual patients. The authors defer to the recommendations of the American College of Physicians, which are accurate but hardly helpful, and omit what The “New” Prostate Cancer InfoLink would consider to be a critical element, which is covered elsewhere in the article. That is the importance of taking account of the patient’s, age, life expectancy, family history, race/ethnicity, and other personal health factors in making the decision whether testing for prostate cancer is appropriate or not.
  • The article does not discuss, at all, the potential merits of  “baseline” PSA testing (at any specific age). We consider this to be a pity.

The editorial by Brawley and Boyle is an opinion piece, and it is more opinionated than the main article. Furthermore, in our opinion, it uses some “loaded” language in making the correct recommendation that regular, mass, population-based screening is not currently justified based on the available evidence. That “loaded” language is centered around the use of the terms “over-diagnosis” and “over-treatment.”

So let us be very clear. First and foremost, there is excellent evidence today (see a recent article by Keen and Keen) that “mass, population-based screening” using mammograms to look for breast cancer is no more justified than prostate cancer screening, on any good statistical basis. We look forward with excitement to the publication of an editorial by Dr. Brawley (in his capacity as chief medical officer of the American Cancer Society) and a specialist in breast cancer epidemiology that makes that point in an American Cancer Society publication, together with the corollary that breast cancer is “over-diagnosed” and “over-treated.” In their article, Keen and Keen state that “2,970 women must be screened once” to find 27 cancers and save one life (in women aged between 40 and 65 years of age). In Brawley and Boyle’s editorial, they repeat the finding of the ERSPC screening trial that it would be necessary to screen 1,410 men and find an additional 48 cancers to prevent one prostate cancer-specific death.

Second, “over-diagnosis” and “over-treatment” are (at least to some extent) a consequence of “over-reaction” on the part of those who make their living from the management of urological malignancies — and their patients. Now there are many good reasons for this behavior. The physicians are, in America, faced with significant legal risk if they do not suggest treatment. And patients have been taught, over the past 50 years, that cancer is a deadly disease: they want it “out.” But there are simple answers to this issue of over-reaction, and they start with greater honesty — among the clinical community and among the survivor community — about what we really do and don’t know.

Over the past 30 years, in America, we have come from a place in which nearly all prostate cancer was diagnosed in men with locally advanced or metastatic disease to a place in which 80+ percent of prostate cancer is diagnosed as apparently localized and potentially organ-confined disease that can be treated initially with curative intent. This has led, for whatever reason, to a reduction in mortality from around 35,000 prostate cancer-specific deaths per year to about 28,000 — a 7,000 reduction in numbers of deaths per annum. There has been a cost associated with this reduction in mortality. That cost is measured both in dollars spent and in the trivial or major adverse effects associated with treatment in men who may never have needed treatment at all. But we still can’t tell who they are beforehand, and so fear and “standard practice”  tell us that we should proceed with treatment “to be on the safe side.” We need to do better. And it doesn’t help to demonize the problem with terms like “over-diagnosis” and “over-treatment.”

Given all of that, Brawley and Boyle still and accurately make the points that:

  • “Shared decisions to use or not use PSA testing for the early detection of prostate cancer should remain within the physician-patient relationship, and should include discussion of the … risks and benefits.”
  • “Shared decision making, compared with simple ‘informed consent,’ should become standard.”
  • “[T]he term ‘shared decision making’ ” is used “to stress that the weight of the decision should not be thrown into the patient’s lap.”

Why do they make these points? Because, as already indicated above, in all truth we do not have good enough information to allow us to know the best thing to do for the vast majority of men who are at only a statistical (as opposed to a clinically evident) risk for prostate cancer. And 28,000 men will still die in America of prostate cancer this year.

8 Responses

  1. First of all thank you for giving your in-depth view. I’ll have a more thorough read later.

    I just hope the USA doesn’t over react and go the other way regarding the two initial tests, especially when read along with this article that states:

    “Until now, if a man could not decide whether he should have his PSA checked, the default recommendation was in favor of testing. These studies suggest the opposite strategy might be better — that unless a man has a particular reason to request a test, the default should be to skip it. ”

    You could end up with 56k deaths a year!

  2. In the article, Brawley et al. mention:

    “The decrease in prostate cancer mortality, occurring just 4 to 5 years after screening began, is difficult to attribute to screening when dealing with a disease of such a long natural history and supports alternative explanations for the decline in mortality. During the same period as screening proliferated, significant improvements developed in surgery and radiation as well as in the application of hormonal therapies to regional and metastatic disease. One such change was the increased use of gonadotropin–releasing hormone agonists and the use of oral androgen–blocking drugs. The dramatic increase in the number of localized prostate cancers detected also simply provided physicians with a greater opportunity to hone treatments, increasing their curative nature while focusing as well on reducing toxicities.”

    This is risible because without PSA testing all these advances in surgery, etc., would be invalid. The fact that a stage shift occurred should not be ignored. Typically, treatments applied at an early stage result in better outcomes. The 7,000+ lives and counting saved every year still remain unexplained.

    Mike, I am willing to bet a nickel that you’ll never hear Dr. Brawley address how many women have to be tested to save a life. BCa is the blood life of the ACS. He will never go there ….

  3. To Do…

    Using human males instead of mice, determine once and for all whether it is possible to intercede medically with men who are diagnosed with prostate cancer, if such diagnoses ever become reliable and available, and how to do so in an expeditious, inexpensive, humane manner for males of the human species everywhere, and to quit screwing around with naked mice.

    Phil Olsen, HPPC

  4. Is there a bit of conflict between the authors of the articles and the AUA, just out of interest?

    I thought April’s AUA “Best Practice Statement” and the section on “The Use of PSA for Early Detection of Prostate Cancer” was as near spot on as you can possibly get with the two initial tests we have today.

    There was no mention of this statement within the two articles which I find strange and as earlier stated no mention of gaining a “baseline” figure to aid future assessment which is not just a “pity” but an unbelievable omittance, in my opinion.

    Overall the two articles are more or less inclined towards the AUA “Best Practice Statement” if you remove the “Baseline PSA age 40 years with anticipated lifespan of 10 or more years” part out of the flow diagram on page 9.

    And yes, a new test is required and will arrive but for now we have to use the two initial tests, adjusting them to both catch prostate cancer as early as possible along with avoiding unnecessary evasive treatments while at the same time refraining from emotive language and statistics to confuse the issue further!

  5. I cannot accept the words “over-diagnosing.” By diagnosing, a cancer has been found. That is important information. The problem of “over-treating” falls into the responsibility of the physicians who are performing the “over-treating” when diagnostics indicate that active surveillance (AS) would be a reasonable option. Discussing AS with patients, when diagnostics dictate this a reasonable option, should be a physician responsibility, but has rarely been done in the past. If the patient then opts for treatment rather than AS, that treatment should not be considered “over-treatment.” It is rather patient choice and should not be included in statistics used to make the case for “over-treatment.”

  6. It is somewhat satisfying to me to read this article and the responses to this article as it provides evidence that I am not alone in being upset with the current miserable state of lack of credible information and lack of effective screening tests for prostate cancer.

    As a patient, I find it ridiculous that we are in this dilemma and the current best advice that our doctors can provide is “Let’ s make a shared decision,” which is about as effective as both the doctor and patient throwing a dart at a dartboard to seek answers.

  7. Unfortunately the publicity given to this publication and the editorial that accompanied it does not add to an open and realistic dialogue about prostate cancer and the problems that men and their doctors have to deal with on a daily basis. I wonder if the timing of this article and the early surprise release of the PLCO trial results is tied to the discussion of health care reform?

    I am concerned that decisions will not be made based on the good of the individual. I am upset because, at least based on what Mike said above, in the editorial there was little consideration to a physician doing a differential diagnosis. Seems to be an attitude of all or nothing. Too bad.

  8. I ask the American Cancer Society and Dr. Otis Brawley to offer alternatives to the estimated 28,000 or more American men who ACS estimates will die of prostate cancer annually.

    Castigating the only widely available test, the PSA test, as the source of the problem is less than helpful, and deliberately misleading.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: