The prostate cancer screening controversy: an evolving, risk-based perspective

WARNING: This commentary may not be suitable reading for those who are firmly convinced that annual, mass, population-based screening of all men over (say) 40 or 50 years of age is the right way to manage prostate cancer risk.

A Reuter’s report just over a week ago was headlined, “No evidence for prostate cancer screening: study.”  The full text of the article by Djulbegovic et al., on which this statement is based, appeared in a recent issue of the British Medical Journal, together with the full text of another article, by Vickers et al., on the use of PSA testing to predict 25-year risk of clinically significant prostate cancer in 60-year-old Swedish males, and full text of an editorial commentary by Andriole on both articles.

The “New” Prostate Cancer InfoLink is of the opinion that these three articles (along with other data that we intend to refer to) offer us all some insight into a potentially better way to think about managing risk for prostate cancer in the US — and in other populations. We fully appreciate that the views we shall express in this commentary are not going to win us any popularity awards in some sectors (hence the introductory comment), but  it is the duty of the good advocate to think hard about what is “right” and ask tough questions about “perceived truth.”

Let us begin with the paper by Djulbegovic et al.

In this paper, the authors provide a detailed discussion of a meta-analysis of data from six, large, recently published, randomized prostate cancer screening trials (in Canada, the US, and Europe) that included a grand total of 387,000+ participants. It is true that there are all sorts of problems with all six of these studies, and it is also true that the follow-up in these studies is not long enough to establish “the truth” about whether mass, population-based screening for prostate cancer is able to provide some effect on overall or prostate cancer-specific mortality at (say) 15 or 20 years of follow-up. However, it is also a fair question whether we should wait another 10 years or so to establish that “truth” when the only one of the six studies that did, in fact, demonstrate a clear prostate cancer-specific mortality benefit (the Göteborg study) did not demonstrate an overall mortality benefit and was conducted in a previously untested population.

We are not going to go into a detailed discussion of this meta-analysis. The data are available for everyone to read, and Andriole (the lead investigator for one of the six trials discussed) makes some highly pertinent comments about this analysis in his editorial. The authors themselves reach three fundamental conclusions:

  • The “existing evidence from randomized controlled trials does not support the routine use of screening for prostate cancer.”
  • However, “screening probably aids in earlier diagnosis and helps to detect prostate cancer at an earlier stage.”
  • “This early detection, which has not been shown to have a significant impact on mortality, comes at the price of additional testing, the risk of overtreatment and downstream adverse effects, and impaired quality of life that currently cannot be precisely quantified.”

(Please note that the authors do not use the perjorative term, “over-diagnosis” in drawing these conclusions!)

Let us now move on to look at the paper by Vickers et al. This paper needs to be considered as a complementary paper to two earlier articles —  published by Lilja et al. in 2007 and by Umert et al. in 2008. In the first of these three papers, Lilja et al. demonstrated that it was possible to predict 25-year risk of prostate cancer diagnosis in a well-defined cohort of > 20,000 Swedish males based on blood samples taken between 1974 and 1986, when these men were between 44 and 50 years of age. In the second paper, by Umert et al., the authors used data from the same cohort of > 20,000 Swedish males to evaluate their 25-year risk for clinically significant prostate cancer. In the newest paper, Vickers et al. have shown that it is also possible to predict the 25-year risk for prostate cancer metastasis and prostate cancer-specific mortality in another well-defined cohort of Swedish males based on blood samples taken in 1981 when all of the men were 60 years of age.

Is there a simple take-away from these three Swedish studies that correlates to the data from the meta-analysis of the six major screening trials? We believe that there is, and it is based on a testable premise:

  • All men should get a PSA test every 5 years starting at age 40, each of which is likely to be able to project a 25-year risk for diagnosis of prostate cancer and/or clinically significant prostate cancer.
  • Men who are shown to be at no significant 25-year risk based on these 5-yearly PSA tests may not need to get interim PSA tests unless there are other reasons for them to do so (based on clinical signs and symptoms, ethnicity, genetics, and other known factors).
  • Men who are shown to be at significant 25-year risk based on any one of the 5-yearly PSA tests should be encouraged to monitor their PSA with care over time in consultation with a prostate cancer specialist and make appropriate clinical decisions based on their individual data.

These suggestions alone will not eliminate risk for over-treatment because over-treatment can still occur based on overly aggressive use of biopsies and subsequent treatment of indolent forms of prostate cancer. What these suggestions do do, however, is define a testable platform for prostate cancer risk management into the future — one within which we use better defined risk factors to make better decisions about the real risks for clinically significant prostate cancer. In addition, these suggestions leave in place the idea that some men with low PSA levels are at greater fundamental risk than others, and therefore may need to think about PSA testing more often than every 5 years to manage that risk.

We find it particularly interesting that back in 2005, Aus et al. proposed, on the basis of data from their own prospective, randomized study of early detection for prostate cancer in > 5,800 Swedish males, that “Retesting intervals should be individualized on the basis of the PSA level, and the large group of men with PSA levels of less than 1 ng/ml can safely be scheduled for a 3-year testing interval.”

Of course we still need a much better test for risk of clinically significant prostate cancer, but, until such a test is available, we believe that the last word should be left to Dr. Andriole, whose editorial comments in the British Medical Journal conclude as follows:

For now, clinicians are best advised to individualize their approach to PSA based screening. Young men at high risk of prostate cancer, such as those with a strong family history and higher baseline PSA concentrations, should be followed closely and could also be considered for ‘risk reduction’ approaches with 5α-reductase inhibitors or dietary and lifestyle modifications (or both). Conversely, elderly men and those with a low risk of disease could be tested less often, if at all. Approaches such as these will hopefully make the next 20 years of PSA based screening better than the first 20.

20 Responses

  1. Not good news for the thousands of younger men in the coming years that can be diagnosed while the disease is still weak. My growing list of young men diagnosed with prostate cancer will never agree with this post.

    We do need that better test, but we also need a happy median for those who are unnecessarily treated for prostate cancer. I still believe that education trumps skipping the screening. I am severely concerned that we are running the risk of losing the attention needed to draw research funding to find that test.

    But that stated, we are definitely headed backwards 30 years in many ways. I just have to get used to it. …

  2. Tony:

    You and those “thousands of younger men” would still get appropriately diagnosed early under this scenario. Indeed, there’s a fair chance you personally would have been diagnosed at 40 or 41 instead of 44. I wil make a bet that your PSA was significantly > 1.0 ng/ml when you were 40, which would have placed you at significant risk and suggested you needed careful annual monitoring.

  3. I thought your opening warning was reminsicent of one I used on my posts some 10 years ago when I started suggesting that perhaps it wasn’t essential to treat every diagnosed case of prostate cancer as if it was life threatening. Here’s what I headed my posts with:

    HEALTH WARNING: This post may contain information or views which may jar your current paradigm. Should you wish to stay in the tranquillity of your comfort zone hit the Delete button NOW. Proceeding to read this post implies acceptance of the fact that you are prepared to possibly learn something new or see existing information from a different viewpoint.

    I didn’t use it for long — being threatened with physical mayhem made me realise that not everyone shared my sense of humour.

  4. Maybe so. Unfortunately, such a plan was not in place 4 years ago for the younger guys. I was in a routine physical when my primary care physician checked the PSA box without even telling so. Perhaps even this would be frowned upon today …

    In one of the support groups I moderate we had the wife of a man posting about the growingly desperate situation her young husband was in. He had been screened at age 45 as that was the appropriate age at the time for a man with family history, and his PSA was well within safe limits. His next test was at age 50 and his PSA was above 100. About a year later the sad post from his wife reminded us that prostate cancer is not such a predictable disease after all. We had lost another young man to it. Even the above plan would likely have been an insufficient plan to find an opportunity that could have saved this man’s life.


  5. I’ll start with your suggestion:

    • “All men should get a PSA test every 5 years starting at age 40, each of which is likely to be able to project a 25-year risk for diagnosis of prostate cancer and/or clinically significant prostate cancer.
    • “Men who are shown to be at no significant 25-year risk based on these 5-yearly PSA tests may not need to get interim PSA tests unless there are other reasons for them to do so (based on clinical signs and symptoms, ethnicity, genetics, and other known factors).
    • “Men who are shown to be at significant 25-year risk based on any one of the 5-yearly PSA tests should be encouraged to monitor their PSA with care over time in consultation with a prostate cancer specialist and make appropriate clinical decisions based on their individual data.”

    That would likely have worked for me. I wasn’t diagnosed until late in my 59th year, and by that time it turns our my gland volume was 101 cc — exceptionally enlarged – though I had no idea and wasn’t aware until my prostate was surgically removed. The urologist never mentioned that huge a gland nor offering BPH-reducing drugs when suggesting RP. In any event, turns out by then, and with a PSA of 6.8 ng/ml, a tumor had extended into “fatty tissue” so I had to also go through the series of EBRT. And despite lymph nodes, seminal vesicles, and vas deferens clear of cancer, my cancer recurred 3 years later. If I had been on either a nationally mandated or recommended PSA testing and DRE beginning at age 40, and then every 5 years thereafter, my cancer would likely have been caught in much earlier stage and possibly eradicated.

    The concern I have, however, is for those men now well beyond age 40 receiving no recommendation for even annual PSA testing and DRE examination. So, let’s encourage a national recommendation to begin PSA testing and DRE beginning at age 40, and if nothing untoward is evident, then every 5 years. But for those men currently 50 and over, let’s continue the recommendation for early attention to PSA testing and DRE examination and then determine how often these men should continue to be tested. Not just ignore this importance. And get the urology and radiation organizations to provide guidelines for their physician members to explain active surveillance and avoid recommending immediate treatment when not warranted by diagnostics.

  6. Tony:

    The Lilja and Umert and Vickers papers base decisions on PSA levels way below the “safe” limits of 3 or 4 ng/ml. We are talking about predictions based on PSA levels of 0.5 to 1.0 ng/ml dependent on patient age. We need to find a way to move forward.


    Actually the Lilja, Umert, and Vickers studies, when considered as a group, imply the possibility of 25-year risk projection even for men who only start getting PSA tests at age 50 or 55, so regardless of when you start getting tested, it should be possible to make good decisions about whether the patient is at risk (i.e., has a PSA above some level like 0.5 or 1.0 ng/ml) or is not at risk. In the former case, more regular PSA tests would be appropriate. In the latter, the patient arguably doesn’t need another test for 5 years. Our first goal should be to stop testing (any more than we have to) the 5 of 6 men who are at no risk for prostate cancer. The second, as you rightly observe, is to stop treating the 30-40 percent of men who don’t need treatment because they have indolent disease.

  7. Don’t get me wrong, I like the recommendations for the most part. I imagine Otis Brawley will have to adjust more than I will. I just have seen sudden turns in men who would fall into the assumption category that a 25 year risk can be predicted. And if you wanted to bet that if I was tested at age 40 I could have been detected earlier or even at all, you may or may not win that bet. Remember I live in Las Vegas, I could have probably found you a house willing to take that bet. :-)

  8. Until we have a better test, there are always going to be issues simply because the PSA test is not cancer specific. And even when we get a better test, it still won’t be 100% accurate for all men. That’s lot the way life works. … Oh, but you know that don’t you. You live in Las Vegas!

    Can you get odds on the outcome of a radical prostatectomy? Maybe we need a sideline business for Prostate Cancer International as a fundraising mechanism!

  9. My fear is that the man you (the doctor) test at 5-year intervals might be the 1 of 6 whose PSA rises quickly but asymptomatically soon after year 1 or 2 before you routinely screen him again that 5th year. That possib-lity, however remote, makes me wonder whether you might be gambling with his life, whether you’re in Las Vegas or elsewhere.

  10. Mike,

    That would be what is called in Vegas a proposition bet. Or noted by the high rollers ~ a sucker bet.

    At this point I know a great new test would likely only help those behind me. I need something better than a better test, I hope for a cure. I need a treatment that trumps all detection tests. That would unfortunately put PCaI, and many others, out of business …

  11. Dear Rabbi Ed:

    I don’t know where you are getting any statistic that suggests 1 man in 6 who is going to get clinically significant prostate cancer has a PSA that rises that rapidly within that short a time period from a level of 1.0 ng/ml or less (unless they have other known risk factors which would already suggest more frequent testing). Can you provide me with a reference?

    I recognize that there may be a very small percentage of men for whom that might be the case, but according the Lilja and Vickers papers this percentage of men is certainly small. In the paper by Vickers et al., of men aged 60 with PSA concentrations of ≤1 ng/ml the risk of metastasis was 0.5% (5 in 1,000) by age 85 and risk of death from prostate cancer at age 85 was 0.2% (2 in 1,000). However, based on the protocol outlined above, even these men should get further PSA tests at 65 and 70 years of age and would have a high chance of being identified early in the disease state.

  12. Tony:

    Let me be extremely clear. We actually need three separate but interlinked items:

    (1) We need a test that can accurately identify clinically significant prostate cancer with very high specificity and selectivity (and ideally without the need for a biopsy).

    (2) We need curative therapy (or therapies) that have minimal or no short- or long-term side effects.

    (3) Once we have these, we can then — with my full cooperation and support — put PCaI and others “out of business.” I can think of lots of other things to do! (Although PCaI’s co-founder, Dr. Krongrad, would definitely need to find a new source of income!)

  13. We did have one of those men who regularly got annual (not 5-year) testing with normal results. But he missed the test one year. It must have been shortly after that “miss” that his prostate cancer became exceptionally aggressive, since the year after missing the test — thus a 2-year interval of aggressiveness — his cancer was Gleason 9 and metastasized to several areas of his body. He opted for orchiectomy figuring shutting down that source of testosterone would be more effective than LHRH agonists and antiandrogens, but to no avail. His cancer kept the aggressiveness, became hormone refractory, failed every chemotherapy agent, and he died within 5 years with his system, under imaging, lighting up everywhere. So, in this man’s case, waiting 5 years after the last “regular” testing, he would have already been consumed with cancer and likely already passed on even earlier. There are certainly going to be exceptions wherein even the “every 5 years” after a normal PSA check and DRE examination will not be adequate. It just makes little sense to even put aside the annual PSA testing and simple DRE examination. This appears to me to have all come about because of the “studies” that argued that men were being “over-treated,” thus appearing to be a “cost” factor to the insurance companies to bring up this subject. And that “over-treatment” came at the hands of urologists and radiation oncologists encouraging “early treatment” to even low grade, near-indolent disease.

  14. In my last post, rather than saying his cancer got aggressive shortly after that “miss,” I meant to say that because of that miss, and his cancer likely becomming aggressive shortly after his last actual test, that miss resulted in the aggressive cancer having two years to do its damage before being recognized with his next test two years later.

  15. Chuck:

    Again … I would ask what this patient’s PSA was at his last test before the cancer became aggressive. If his PSA was higher than 1.0 ng/ml at his last test before the cancer got aggressive, based on the Lilja/Vickers data, then he should have been getting annual tests anyway. It is not enough for the patient’s PSA to be “normal” (meaning somewhere below 4.0 — or even 2.5). The Lilja/Vickers criteria for 25-year risk is 1.0 ng/ml or less. And although people may not wish to deal with it, there is an enormous annual cost associated with giving PSA tests to the 5/6 men who are never going to get prostate cancer which has been estimated in the billions of dollars, much of which is getting covered by Medicare and therefore coming out of your pocket (not the insurance companies)!

  16. Now you have piqued my interest in the numbers of men who are found to have PSA levels of 1.0 ng/ml or less to then be re-examined every 5 years. With the medical community already having considered dropping the concern level from 4.0 ng/ml down to 2.5 ng/ml, and the apparent couple hundred thousand men who are being found annually with PSA levels over that more recent 2.5 ng/ml concern level, the necessity of annual testing is once again confirmed. The concern remains that with 5-year spans between examination, how many men will have experienced aggressive prostate cancer development during that span that has not only jumped well over 2.5 ng/ml, but metastasized as well? We have no idea, and it is because we have no idea that the simplicity of annual PSA testing and DRE examination should continue. Breast cancer is a sufficient concern that women are regularly encouraged for examination — not 5-year spans. Prostate cancer should be afforded the same concern. If I have said it once, I have likely said it many more times, the real concern is the physicians (urologists and radiation oncologists) who encourage early treatment when such early treatment may not be required. When a man diagnosed with early prostate cancer development makes his personal choice to want to treat his cancer rather than following an active surveillance program, that is certainly his choice and can not be categorized as “over-treatment.”

  17. Here’s a link to responses to Andriole’s editorial in the BMJ that are interesting.

    My problem with Andriole’s recommendation that docs “individualize” the screening process is that not all docs are the same. Many “Main Street” urologists* screen and treat prostate cancer based on standards of care that are far, far, far, far from “cutting edge.” (Prostate cancer screening and treatment tip: Don’t let your wife take a journal article that the Doc hasn’t read — and possibly hasn’t even heard of the journal, or that there are journals — to an appointment. It will only make the doc defensive and entrenched in his ignorance and your wife won’t stop crying and bitching until you transfer to a doc with a clue.) When they do introduce new protocols into their practice, it’s for purposes of their “business model” — like investing in a da Vinci — rather than a desire to provide the best possible care with the best possible outcomes for the individual patient.

    Remember the old saw about the guy looking for his keys under the street lamp though he lost them in the woods? Most men — particularly those in medically under-served communities, those who are less affluent, and those with lower levels of education — are treated by Dr. Lost-his-keys.

    The “every man for himself” recommendation may work splendidly for men fortunate enough to have access to the Andrioles of the world, but it’s a dangerous proposition for millions of men who lack access to docs with the knowledge, wisdom, time, and interest to treat them “individually.”

    Then there’s the bigger policy argument about limiting screening to expensive-to-treat diseases to worry about. Health care providers making health care policy like to play the “suspension of disbelief” game. They try to convince us and themselves that money doesn’t matter. This is a dangerous game. Regardless of the medical efficacy of any particular screening or treatment, we have to factor in the desire of the health insurance industry to profit, profit, profit — first, foremost, and always.

    Perhaps it’s worth the cost of screening to mens’ health and taxpayers (as Mike pointed out) to “over-screen” lest the insurance industry decide to simply not cover any screening or early treatment at all? I’m with Tony from a policy perspective. Any chance the insurance industry has to not provide care, they’ll take it. We simply can’t let substantive arguments about the efficacy of screening protocols be co-opted by those with agendas that have nothing to do with health.

    * This assumes that men are being screened by urologists. The reality is that most population-based screenings are conducted by internists and family practice docs, who — even if they’re meticulous in their desire to keep up with the literature — can’t possibly keep up with “cutting edge” protocols for every condition they treat.

  18. All of the quoted studies present conclusive evidence that PSA testing is an excellent diagnostic tool.

    I strongly support the recommendation to test at 40.

  19. If only one of the above studies showed a prostate cancer-specific mortality benefit for early detection, why are we even debating any of this? It is clear early detection does not save lives.

  20. Chris: Different people see different things in these data because they are not 100% compelling one way or the other.

Leave a Reply

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

You are commenting using your 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: