New study won’t end debate on PSA test for prostate cancer

Robert H. Shmerling, MD

Senior Faculty Editor, Harvard Health Publishing

Whether or not men should have a simple blood test to look for hidden prostate cancer has long been a controversial and confusing issue. A new study from Europe does little to resolve it.

Dr. Harvey Simon, editor of the Harvard Men’s Health Watch, has written thousands of words, and counseled many of his patients, about PSA testing. In the March 2012 issue, he discusses his own decision not to have routine PSA testing.

In this study of 162,000 men, those tested for prostate-specific antigen (PSA) were less likely to have died of prostate cancer over the course of the 11-year study than men who didn’t have the test. Deaths from any cause, however, were the same in both groups—meaning that the test didn’t save lives.

The researchers (and most news reports) highlighted the 21% relative reduction in deaths due to prostate cancer. That sounds like a lot. But here’s what it means in absolute terms: among men who had the PSA test, there were 4 deaths per 10,000 men per year, compared to 5 deaths per 10,000 men per year in the group that didn’t have the PSA test. The study was published in the New England Journal of Medicine.

What can be bad about a test for cancer?

On the surface, a relatively inexpensive test that can detect cancer in its earliest (and most treatable) stage looks like a winner. But the devil is in the details:

  • Many men with prostate cancer are elderly and have no health problems related to the cancer. They may have other, more important health problems. For them, early detection of prostate cancer usually isn’t helpful.
  • Prostate cancer often grows very slowly. Most men who have it die of something else. Some men live with the side effects of treatment—notably impotence and incontinence—for a cancer that would have had no effect on the length or quality of their lives.
  • Since a high PSA does not always indicate cancer, a biopsy is needed to find out why the PSA is high. Prostate biopsy isn’t the most pleasant of procedures and it can be accompanied by complications such as bleeding or infection.
  • Studies have not proven that PSA testing saves lives.

The U.S. Preventive Services Task Force, an independent panel of experts whose recommendations help define high-quality health care for most Americans, has long recommended that men over age 75 not get routine PSA tests unless there is a good reason to do so. Last fall, the task force issued a preliminary recommendation (now awaiting public comment) saying no to routine PSA testing for all men.

Rethinking the PSA test

The new study is likely to strengthen the task force’s advice against yearly PSA testing. My own view is that we should reconsider the widespread use of PSA testing, especially the yearly screening that is common in the United States. To be worthwhile, a screening test should have clear benefits that strongly outweigh any risks. That hasn’t been proven for PSA testing.

This makes PSA testing a most individual decision. As my colleague Dr. Harvey Simon writes in the Harvard Men’s Health Watch, all men should talk about PSA testing with their doctors. Learn the risks, benefits, and limits of this test. Then decide with your doctor whether you should have the PSA test at all and, if so, how often.


  1. jeryy-00

    Never any good news, always doom and gloom. When will ya’ll be satisfied, when lettuce is 10 bucks a head and milk 20 bucks a gallon. Its always about a tiny fish or a breath of fresh air, NEVER the human race. Can’t clear, can’t fertilize, can’t spray, can’t water, can’t use fuel, how do you want a farmer to produce anything, or do you. GET REAL

  2. jojolendir

    great article…i have to admit

  3. Bob D

    I have a small business that employs some migrant workers and it seems this topic does not get discussed much in the hispanic community, It would be helpful to reach out to this community

    Bob D

  4. Tim Bartik

    From a scientific point of view, this article and other coverage of this issue should not put any stress whatsoever on the European study’s finding that there are no statistically significant effects on all-cause mortality. This study’s sample size was simply much too small to detect a decrease in all-cause mortality of the size of the reduction in prostate cancer mortality in this study. Actually, the point estimate is that all-cause mortality actually decreased in the screening group, relative to the control group, by MORE than the reduction in prostate cancer mortality. But this decrease is statistically insignificant because the baseline for all-cause mortality is about 20%, and detecting a reduction from 20% to 19.9% is impossible with this study’s sample size. If you do a simple power analysis, the sample sizes of the European study only have a power of 8% to detect a reduction in all-cause mortality equal to the study’s estimated reduction in prostate cancer mortality. To have a study with a power of 0.80 to detect a decrease in all-cause mortality of this size would take a study with 2.3 million in the screening group and 2.3 million in the control group. Therefore, the lack of statistical significance to the estimated effects on all-cause mortality is a red herring, a distraction from the fundamental issues. No such statistically significant effects on all-cause mortality could ever have been reasonably expected from the design of the European study.

  5. John Williams

    I understand that there is a concern over the validity of PSA testing, and I understand that 21% may in fact represent a small percentage.

    But when you are suffering the possible effects of prostate cancer, urgency to urinate, inability to hold on to your urine, the consent embarrassment and pain associated with this condition, then a inexpensive test is of invaluable tool that helps relieve the stress and uncertainly of this irritating issue.

    And ‘Yes’ I do suffer from BHP.


  6. gabe

    Stewart Justman wrote an e-single on this topic–the emphasis on screenings/early detection of prostate cancer, and how it can lead to over diagnosis and end up hurting patients. Very interesting read:

    • Nurwahidah

      There is nothing to penvert you getting any kinf of cancer. You can try to lower the risks but at the end of the day they say if its in your genes and you have that particular virus, you are going to get it regardless of what you do.

Commenting has been closed for this post.