PSA – To Screen or Not to Screen

You’ve all heard the dramatic testimonials in the media: “I had a PSA test and they found my prostate cancer early enough to treat it. The test saved my life. You should get tested too.” The subject of screening tests is one that confuses the public. On the surface, it would seem that if you can screen everyone and find abnormalities before they become symptomatic, only good would result. That’s not true. Screening tests do harm as well as good, and we need to carefully consider the trade-offs.

About half of American men over the age of 50 have had a PSA (prostate-specific antigen) screening test for prostate cancer. Recommendations for screening vary. The US Preventive Services Taskforce (USPSTF) says there is insufficient evidence to recommend screening. The American Urological Association and the American Cancer Society recommend screening. Urologists practice what they preach: 95% of male urologists over the age of 50 have been screened. But other groups like the American Academy of Family Physicians recommend discussing the pros and cons of screening with patients and letting them make an informed choice.

Two recent studies published simultaneously in The New England Journal of Medicine have added to the controversy. One concluded that screening does not reduce deaths from prostate cancer; the other concluded that it reduces deaths by 20%.

The PCLO Cancer Screening Trial

The Prostate, Lung, Colorectal and Ovarian (PCLO) Cancer Screening Trial involved 76,693 men in 10 US study centers. Subjects in the study group were offered annual screening with PSA for 6 years and annual digital rectal exams for 4 years; the control group received “usual care” (which could include screening – 52 percent of the control group had screening in the 6th year.) They found 2820 cancers in the screening group and 2322 cancers in the control group. There were 50 deaths in the screening group and 44 deaths in the control group, not significantly different. There was no reduction of prostate cancer mortality in 7 years of follow-up. One could argue that the effect of screening was diluted by the high rate of screening in the control group, or that longer follow-up might show a difference.

The European Randomized Study of Screening for Prostate Cancer

In the European Randomized Study of Screening for Prostate Cancer, 182,000 men age 50-74 in 7 countries were offered PSA screening every 4 years. 82% accepted at least one offer of screening. After 9 years follow-up the death rate from prostate cancer decreased by 20% but was associated with a high risk of overdiagnosis. There were 214 prostate cancer deaths in the screened group and 326 in the control group. From the statistics of those who were actually screened, one can calculate that 1068 men would need to be screened and 48 would need to be treated to prevent one death from prostate cancer. There was no reduction in all-cause mortality – what does a reduction in deaths from prostate cancer mean if there is no reduction of deaths overall? The absolute reduction in deaths from prostate cancer was 7 deaths per 10,000 men screened. 73,000 men in the screening group underwent more than 17,000 biopsies.

The Downside

For any screening test, the benefits must be weighed against the risks. There are significant risks to PSA screening: biopsies, surgery, hormone treatment, chemotherapy, impotence, incontinence. Some of the men who suffer these invasive procedures and side effects would probably have lived to die of something else if they had not been diagnosed. Many prostate cancers are slow-growing and not actually life-threatening. For most of these men, it would probably have been better if they had not been screened and had never known they had prostate cancer. Screening leads to over-diagnosis, and some people get treated because of false positive tests – they never actually had the disease.

It boils down to a value judgment. Is it worth screening 1068 men and treating 48 to save one life from prostate cancer when you don’t save lives overall? How do you weigh the costs and the suffering against the value of a life saved? There is no easy answer. Future research will try to assess things like quality of life and cost effectiveness, and we may come up with better recommendations about when to test and what level of PSA to use as a cut-off.

In an accompanying editorial in the same issue of NEJM, Michael J. Barry MD concluded:

PSA screening has at best a modest effect on prostate-cancer mortality during the first decade of follow-up. This benefit comes at the cost of substantial overdiagnosis and overtreatment. It is important to remember that the key question is not whether PSA screening is effective but whether it does more good than harm. For this reason, comparisons of the ERSPC estimates of the effectiveness of PSA screening with, for example, the similarly modest effectiveness of breast-cancer screening cannot be made without simultaneously appreciating the much higher risks of overdiagnosis and overtreatment associated with PSA screening.

Posted in: Clinical Trials, Diagnostic tests & procedures

Leave a Comment (19) ↓

19 thoughts on “PSA – To Screen or Not to Screen

  1. tjdowling says:

    Did this study address the rates of prostate cancer in men whose fathers have been diagnosed with it? Do you think screening in men who have a family history of prostate cancer outweigh the risks? Thanks. -Tim

  2. qetzal says:

    So, if high PSA levels alone are (arguably) an insufficient basis to undergo subsequent procedures (biopsies, surgery, etc.), what other factors should be used to make those decisions? I can guess some of the likely ones: palpable mass on digital exam, family history, actual PSA magnitude. But, are there any good science-based guidelines for how to include those other factors in deciding how best to proceed?

    I ask because my MD is telling me I’m due for a physical, and I expect he’ll want to test PSA. I don’t mind the test itself, since that’s not really what carries the significant risks. It’s the actions that a positive test may precipitate that are the concern.

  3. DrBadger says:

    I think that PSA testing is not as bad as making a rash decision based on it. If you have a high PSA that is stable over several years, you probably have nothing to worry about. Otherwise, I wouldn’t think of a biopsy (or any other action) unless I had significant risk factors (i.e., mass on DRE)

  4. urology-resident says:

    I have some comments on the European Study.

    – Patients got an average of 1.7 PSA tests during the study (in about 9 years). The aggressive cancers have a small window of opportunity for cure, and screening every 4 years will miss some of them.

    – Biopsies in the European study were done using only 6 cores (except in Finland), which is not the new standard of care (now is 10-12 cores). This means that they probably could have detected more cancers (and I admit, overdiagnose a little more)

    – In order to see a difference in mortality you have to wait at least 10-15 years, since prostate cancer is such a slow growing disease. (The same applies to the American study)

    – When you look at the mortality curves of the screened versus the non screened group, they begin to separate around 7-8 years into the study, and continue to do so, implying a greater difference with time.

    My conclusion from these studies: we have to do a better job at not treating the low risk disease and at identifying the high risk patients that will benefit from treatment. We should also continue not to screen people with less than 10-15 years of life expectancy.

    I think the field of Urology is going though a real Science-Based-Medicine challenge…

  5. David Gorski says:

    This reminds me of the issues about screening for breast cancer:

    The early detection of cancer and improved survival: More complicated than most people think
    Early detection of cancer, part 2: Breast cancer and MRI
    Do over one in five breast cancers detected by mammography alone really spontaneously regress?

    Unfortunately, screening for cancer or other diseases is a complex subject, and it’s often hard to demonstrate an actual survival advantage in the screened population.

  6. urology-resident says:


    Risk factors where not taken into account on the European Study as far as I could tell from reading the paper.

    Family history can be tricky: is not the same to have a father that died of prostate cancer when he was 89 than one who died at age 55. In the first case, I would not consider that a significant risk factor but I would on the second.

    I theory, screening high risk population (family history, African Americans) would give more yield but I don’t think this has been proven.

  7. Harriet Hall says:

    Current guidelines for screening are for the whole population and do not address individual factors like family history. I think any risk factors like family history should weigh in the decision to screen an individual. If there is a palpable mass, then PSA is no longer a screening test but part of the diagnostic workup. And the higher the PSA, the greater the risk. The study authors themselves pointed out that longer followup might show benefit.

    Isn’t it ironic that critics of science say it is dogmatic? Here’s a case where the best science we have only adds to our confusion.

  8. hatch_xanadu says:

    Thanks for the post, HH. This is an issue that has been in the mind of the medical community for a while, but I don’t think the general public has been primed.

    Recently there was a bit of a stir about CMS’s decision to not cover CT or “virtual” colonoscopy as a screening tool. As far as I know, the final decision won’t be made until next month, and CMS is taking public comment in the meantime. The essential argument from many proponents of CTC screening was, obviously, that it is an ostensibly less invasive test than colonoscopy and that (and this I thought was the most compelling argument) certain populations, such as African-American males, who are generally recognized as underserved for colon screening, might be more inclined to undergo screening if it didn’t involve an invasive test.

    CMS, however, made an excellent case in their decision summary, taking into account the different stakes involved with screening procedures than with diagnostic or therapeutic procedures, since screening is performed on asymptomatic individuals — and noting that, if polyps are found, the patient then must be referred for optical colonoscopy anyway. In addition, they cited studies that indicated the incidence of small polyps may increase with age and may even approach “normal” for an older population, “more representative of the Medicare population”. And what, then, do you do with, say, an asymptomatic 85-year-old with polyps? I must say that, despite the compelling pleas on the part of CTC proponents, I stand by CMS’s decision.

    Same goes for incidental findings. It’s something medical professionals have been mulling over for some time, but something I don’t think has quite entered the public mind — not to the extent that they appreciate the physician’s dilemma, in any case. Thanks for addressing the dilemma.

  9. James Fox says:

    I wonder how many PCP’s will communicate the ambiguous findings of these studies to their patients so they can make an informed decision as opposed to just sticking with ‘SOP’s’.

  10. Harriet Hall says:

    Clarification. I said the ACS and the AUA recommend screening. They currently recommend “appropriate” screening rather than indiscriminate universal screening of all men.

    Here are the most recent recommendations of the American Cancer Society

    and the American Urological Association

  11. Newcoaster says:

    PSA screening is not paid for by the publically funded health service in British Columbia., though I think it is in most provinces in Canada. (Health care money is provided by the federal government, but the provinces decide how they will spend/manage it) I do have a lot of issues with the bean counters that work in our Health Ministry, but I happen to agree with them on this one.

    It does make for an interesting discussion when men come in and say they want a PSA, assuming that like every other medical service in Canada, it is covered. Of course, they are hoping to avoid the “finger up the bum test”. I tell them getting a PSA won’t avoid me doing a FUTBT. I do try and explain about false positives and negatives, about sensitivity, specificity etc. I’m not sure how many people understand that, but most choose not to pay the extra $30.

    Another problem I’m running across recently is women coming in demanding a CA19-9 to screen for ovarian cancer. I gather there’s an email campaign/chain letter going around that tells women they have a “bad doctor” if he/she won’t order this test.
    It is covered under the health plan, but is also fairly expensive, and I try to do my bit to keep health costs under control by not ordering unnecessary tests. The discussion that I have around PSA screening doesn’t work as well, because they know the test is covered, and often the family docs role as “gatekeeper” is not appreciated.

  12. overshoot says:

    Interesting timing, Dr. Hall.

    I was just in for my annual, and the (call; regular PCP was out) MD and I had a brief chat on the subject. I’ve been getting the PSA and “greasy finger” tests for a decade now [1] and if nothing else that provides a baseline for PSA rate-of-change indications. I’m not out-of-pocket for it, but feel slightly guilty about using resources that arguably could be better used elsewhere. $DOC is a big believer in screening.

    There are significant risks to PSA screening: biopsies, surgery, hormone treatment, chemotherapy, impotence, incontinence.

    I don’t think you meant that as written. The screens don’t themselves have those risks, although they certainly can activate decisions which leads to them. The problem I have with the arguments against screening is that they amount to pleading in favor of ignorance — and I’m rather biased in favor of being informed.

    Then again, I’m not real big on the “ZOMFG! Do something!!!” side of things [2]. There are a lot of conditions where I’m in the “well, them’s the breaks” school. If I start showing a rise in PSA, I’ll see how it looks next year. Time enough to think things through.

    [1] Middle age: when “small and soft” becomes a good thing.
    [2] I’ve been a candidate for knee replacement for at least ten years. We’ll see about that when I can’t ski any more on the OEM version.

  13. delaneypa says:

    Thanks again for a thoughtful article. An experience with a colleague has dissuaded me from employing the thoughtful approach you and profession organizations recommend. Unfortunately, if ever doc in town is checking PSAs willy-nilly, that defines the “standard of care”.

    Please see JAMA. 2004;291:15-16 for a doctor who did the thoughtful long discussion approach and got in trouble. Some bits for those without JAMA access:

    When the trial started on June 23, 2003, I was nervous but confident. I realized that the patient was going to say we had never discussed prostate cancer screening but since I always do and had documented it, I didn’t think this would be a very strong plaintiff argument. What I didn’t anticipate was that the plaintiff’s attorney was going to argue that I should have never discussed the risks and benefits and should have just ordered the PSA. But he did. In fact, a major part of his argument was that there is little risk involved in performing a PSA and that the standard of care is to order the test. Although we had the recommendations from every nationally recognized group supporting my approach and the literature is clear that screening for prostate cancer is controversial, the plaintiff’s attorney argued otherwise…

    As the trial progressed we presented national experts who discussed the controversy surrounding prostate cancer screening and explained some of the potential dangers of PSA. We discussed such things as false positives, indolent vs aggressive cancers, sensitivity and specificity. Our experts explained that because of the questionable benefit vs associated risks of PSA screening, a shared decision by the physician and the patient was recommended by all of the national health associations. The science was clearly in our favor…

    A major part of the plaintiff’s case was that I did not practice the standard of care in the Commonwealth of Virginia. Four physicians testified that when they see male patients older than 50 years, they have no discussion with the patient about prostate cancer screening: they simply do the test. This was a very cogent argument, since in all likelihood more than 50% of physicians do practice this way. One may have argued that we were practicing above the standard of care, but there is no legal precedent for such an argument.

    As is well documented in the literature, physicians take quite a long time to change their patients’ protocols. Thus, we know that many practicing physicians are not using well-proven interventions or implementing well-publicized national guidelines. The legal definition of standard of care protects these physicians and encourages them to change slowly, if at all. It is often claimed that malpractice is a mechanism for holding physicians accountable and improving the quality of care. This case illustrates quite the opposite: punishing the translation of evidence into practice, impeding improvements to care, and ensconcing practices that hurt patients. In our legal system, the physicians who are slow to change are the winners.

  14. Robert L Smith says:

    First of all the studies are already out of date and they haven’t been following up long enough and of course there is the contamination. It’s not screening that causes overtreatment, it’s lack of education, and giving in to fear of cancer. The same fear that leads many to avoid doctors until it’s too late, also leads many to prematurely jump to too much treatment. But don’t through the baby out with the bath water. Teach men men to proactively seek both screening and information as to what the results mean (or more properly the results of a biopsy if that is indicated by the screening). Overtreatment is a statistical fact, but meaningless on the level of an individual since a priori it is difficult to predict the course of the diagnosed cancer. More importantly I certainly hope that each of us as individuals (not faceless numbers in a massive study) needs to be able to utilize our own goals re the risks we want to face. We need to be able to realistically know the odds, the possible outcomes, and attach our own weights to the various possible side effects of treatment as well as the danger of invasive cancer. Life expectancy is by far not the only measure — pursuit of life quality is the fundamental goal and will hopefully remain an intrinsic “right” of each of us interacting with a health care systems that is increasingly intent on grinding us into an undifferentiated statistical mass with a weight in dollars attached to us.

  15. Pingback: PSA «
  16. yeahsurewhatever says:

    First of all, both of those studies were very badly designed. I get the impression that the PLCO trial people don’t understand what “control group” means. Neither group seems to know what “double blind” means. In the US test they let the “control group” self-select for screening, and in the European test they let the screening group self-select for screening, making both inconsistent.

    Second, the two seemingly contradictory studies can be reconciled simply by supposing that “usual care” in Europe is substandard compared to its US equivalent, or that the ERSPC group had an external incentive to find a significant result (which seems to be the case just going by what the article admits) while the US group did not.

Comments are closed.