PSA screening for prostate cancer: a doctor’s perspective

Marc B. Garnick, M.D.

Editor in Chief,

Yesterday’s announcement that men should not get routine PSA tests to check for hidden prostate cancer is sure to spark controversy for months to come. I believe that the U.S. Preventive Services Task Force (USPSTF) made the right decision, and I commend how it came to its conclusion.

On the surface, rejecting the use of a simple blood test that can detect cancer in its early and still-treatable stage sounds foolish. Cynics have been saying it is the handiwork of a group concerned more about health-care rationing and cutting costs than about health. The decision is wise, not foolish, and will improve men’s health, not harm it.

The word “cancer” usually brings to mind images of a fast-growing cluster of cells that, without aggressive treatment, will invade other parts of the body, damage health, and potentially kill. That certainly describes many cancers. But not most prostate cancers. Most of the time, prostate cancer is sloth-like. It tends to grow slowly and remain confined to the prostate gland, with many men never knowing during their entire lives that a cancer was present. These slow-growing prostate cancers cause no symptoms and never threaten health or longevity. That means many men with prostate “cancer” never need treatment.

Before the advent of PSA testing in the 1990s, some men learned they had prostate cancer because of symptoms such as trouble urinating or persistent pain in the pelvic region. Others were diagnosed when a doctor performed a rectal exam and felt suspicious bumps on the prostate. These digital rectal exams remain an important part of a man’s physical exam since they can spot what are called clinically detectable cancers, for which treatment may still be helpful.

A high PSA level is one way to signal other tests, such as a biopsy, that can tell if cancer cells are present. Unfortunately, it can’t tell the difference between a dangerous cancer that requires treatment and one that doesn’t. For that reason, most men who have an elevated PSA and a biopsy that shows cancer cells in the prostate choose to have surgery or radiation therapy. Many of them don’t need treatment, though, and live with side effects such as impotence, incontinence, and rectal bleeding for naught.

Evidence-based decision

The USPSTF is made up of volunteers from a variety of fields, including internal medicine, family medicine, behavioral health, and preventive medicine. None have financial interests in tests or treatments.

The task force based its recommendation about PSA screening on many studies, but the main focus was on two key randomized clinical trials, the gold standard of medical evidence. One, conducted in Europe, had 11 years of follow up, which is extraordinarily long for clinical trials. It compared the health outcomes of men who were offered PSA screening with the outcomes of men who were not offered the PSA test. The panel looked at clear-cut, measureable endpoints such as death, death from prostate cancer, and rates of infection, impotence, incontinence, and other downsides of prostate cancer diagnosis and treatment.

In the outcome that mattered the most—death—there was no difference in overall mortality among men who had the PSA test and those who didn’t, though there was a small decrease in prostate cancer deaths over the 11 years of follow in the screened population. The researchers calculated that 1,410 men would need to be tested, and 48 additional cases of prostate cancer would need to be treated, to prevent one death from prostate cancer. PSA-based screening slightly reduced the rate of death from prostate cancer, “but was associated with a high risk of overdiagnosis,” the authors concluded.

The second trial was the U.S.-based Prostate, Lung, Colorectal, and Ovarian Screening Trial. After 13 years of follow-up, the cumulative death rate from prostate cancer was 3.7 deaths per 10,000 person years in the PSA screening group and 3.4 deaths per 10,000 person-years in the control group. Again, no difference. In contrast to the European study, and in keeping with the practice of medicine as currently practiced here in the U.S., there was no difference in the death rate from prostate cancer in the screened group compared to controls.

All trials have flaws, and advocates of PSA screening say that these flaws undermine the USPSTF recommendation. Had these flaws not been present, they say, the results would have supported the benefits of screening. This charge has not been borne out. In addition, the critics base their arguments on results from just two of the countries in the European study, which had better results than the other five.

Moving ahead

For years, I have been counseling my patients about the uncertainties of routine PSA testing, and more importantly, the uncertainties around treating certain prostate cancers once they’ve been detected.

Many of my patients have chosen to be treated for their prostate cancer. These decisions were made before we know what we do now about PSA testing and subsequent treatment for PSA-detected prostate cancer.

Going forward, making decisions about having the PSA test in the first place to choosing whether or not to start treatment immediately if cancer is diagnosed will have a very different tone, based upon the approach now recommended by the USPSTF. For my patients who will continue to want the test, I will make sure they know that while PSA testing can detect prostate cancer early, many of the cancers it detects will do no harm.

New tests under development may someday be able to tell a dangerous prostate cancer from an indolent one. In the meantime, as I explained in a recent Scientific American article, the practice of medicine should reflect what current studies show, and our decisions should be based upon evidence and not just our beliefs.


  1. Chuck Barker

    I was recently diagnosed w/prostate cancer after PSA results, a DRE, biopsy, and an ensuing robotic prostatectomy.

    A month later, with only a little incontinence but a great deal of sexual disfunction, my diagnosis shows a PSA of 0.270 indicating some part of a grade 8 tumor remains. Radiation therapy plans are being formulated in September.

    I’m 66, in otherwise good health, and after reading all this “new” information re PSA conflicting opinions, I almost wish I had gone in the “wait and see” direction with radiation therapy IF determined to be of necessity.


  2. Stephen Hoey

    I google, google and google again ” life expectancy metastatic prostate cancer. ” The answer regrettably is the same, the cancer cells will become resistant to the hormone therapy and I will die.
    The bad news came from my oncologists at Dana Farber, two years ago at age 52. If caught early I would not be facing certain death.
    At our charity events I have the opportunity to encourage men to ask their doctor for a PSA. Men WILL now use this as a excuse to avoid the test !!!
    My wife of 32 years, my two daughters, my two brother would strongly disagree with the “panel.” One of the doctors on the “panel” is a pediatrician.
    I attend cancer symposium @ Brigham & Womans Hospital , all the oncologist disagree with this backward recommendation.

  3. Anonymous

    @ This is an excellent posting. You have my vote and I will bookmark this website right now.

  4. mover companies

    Very Great Idea, Hope every thing goes success and thank you for giving good post

  5. mover companies

    I truly enjoy looking through on this website , it holds fantastic blog posts.

  6. Angela

    The study concluded that Prostate cancer is being detected with increasing frequency, and many patients are receiving such treatments as radical prostatectomy and radiation therapy. That’s actually pretty intersting.

  7. George Button

    My PSA went up. I had a bx. It was +. I had a robotic. It had metastasized. My father and 2 maternal uncles died from CA prostate. I’ve had RT. My PSA is undetectable. If I had taken your advice I’d be looking at bone mets by now. No thanks. It’s easy to give advice when you’re not a victim. I prefer being around a little longer. I am 100% continent and things still work ok for being 77 and still practicing. GCB, MD.

  8. R Smith

    From what I have read, prostate cancer does take a long time to develop so long term studies are a must. One the other hand approaches to PSA, to biopsy analysis, and to treatment of prostate cancer have evolved and hopefully improved over the past few decades. So at best the USPSTF studies may show that PSA testing averaged over all practices and all treatments 20 years ago are not very helpful. But advances have occurred based on good science (not faceless statistics). So each man, and each woman who loves a man, needs to ask. Who would you go to for advice about prostate cancer — generalist statisticians with no expertise in prostate cancer or leading experts with lots of experience using the latest techniques. Sure in some cases newer doesn’t mean better, and yes some MDs are just motivated by $$ and ego — so find someone you trust, but don’t in my opinion trust someone who just sees you as another statistic.
    Disclaimer: PSA testing in the late 90’s uncovered an otherwise symptomless agressive cancer in my prostate. Anecdotes based on small samples may not be “proof” but certainly are “existence proofs” of the good PSA testing can do — the successful end of the bell curve. Massive statistical studies based an infinite range of conditions aren’t proof of anything, except that there are an equal number of bad outcomes as good outcomes. What the task force should do, if it wants to be truly helpful and not just massage the egos of internists, is find out how to be on the successful end of the bell curve through appropriate use of PSA testing, MD expertise, and present day technniques.

  9. Tim Bartik

    As in his Scientific American article, Dr. Garnick inappropriately puts emphasis on the all-cause mortality results from the European study. The study was “under-powered” to detect such effects. You would need a sample size of several million men to detect declines in all-cause mortality of the order of magnitude of the observed reductions in prostate cancer mortality in the European study. The study’s sample size was designed to be able to be sufficient to detect effects on prostate cancer mortality, but it was never anticipated that the study’s sample size would be sufficient to detect effects on all-cause mortality. Therefore, the study’s failure to do so is not surprising, and tells us nothing.

  10. Dennis

    I agree with Paul. Continue your stance, sir.

  11. Paul from New Haven

    Bravo, sir. I know yours is a decidedly unpopular stance among urologists.

Commenting has been closed for this post.