Wednesday, July 18, 2012

PSA (again)

     Every time I think that the question of PSA testing has been laid to rest, another outcry arises, sometimes pro and sometimes con. I realize that what we are dealing with here is a medical belief system, and belief systems are notoriously impervious to data and facts. I can recall the brouhaha that arose when one group announced that a clinical study showed no benefit of mammograms to women in the 40 to 49 year old age group.

     Let us begin again by realizing that no one argues about the speed of light, or whether or not Albany is the capital of New York State, or if 7 x 8 = 56. No one usually argues about agreed-upon facts that can be measured. The problem arises when two people  try to fit the same data into two different  heuristic structures, because your theoretical picture will frame and interpret the data you view, as all epistemologists acknowledge. There is even more argument when we try to interpret or predict  the result of an intervention on human behavior: we still cannot agree on whether or not the data shows that teaching about sex education and contraception  in high school encourages high school students to have sex. As a general rule, when the data does not lead to an immediately verifiable conclusion (viz. the question of global warming) there is a tremendous amount of heat generated in the argument/discussion with correspondingly little light.

     The broadest discussion to date about the pros and cons of screening with the PSA is contained in this month's issue of the Annals of Internal Medicine (17 July, 2012) , vol. 157, no. 2. There are three articles: The USPSTF (U.S. Preventive Services Task Force) recommendation against PSA screening, with the proviso that the final decision should be one  of "shared decision making" between the doctor and the patient (pp 120-134), a shorter review which concurs with the recommendation of the USPSTF (pp 135-136), and an equally short review that opposes the recommendations (pp 137-138). I will  summarize the three articles from a strictly scientific point of view ---i.e. based on the data cited,  and then try to illustrate where personal beliefs and projected feelings about the quality of life with and without  treatment  may have crept into the discussion. I will also carefully point out the unknowns in this decision making. And let us not forget that the USPSTF recommendations are only recommendations, and each patient will make a decision that is ego-syntonic with his own medical belief system.

     First a caveat. In my practice I have never seen a wife permit her husband to do "watchful waiting" once a biopsy shows prostate cancer. So if you are a husband who would not agree to any treatment, then don't ever do the PSA test.

     1) The only screening method referred to in all the studies is the serum PSA, and not a rectal exam.

     2) More men die with prostate cancer that from prostate cancer. In the latest autopsy study, 70% of men in their 70's have microscopic foci of prostate cancer.

     3) Some groups are at increased risk for prostate cancer: black men and men with first degree relatives who have or have had prostate cancer.

     4) If we define "overdiagnosis" as the detection by PSA of a tumor that will not spread and cause symptoms in the patient's lifetime, then the two largest trials suggest overdiagnosis rates of 17% to 50%.

     5) No study reported on the effect of screening on the development of metastatic disease.

     6) In looking at all the studies, the chance of a non-screened man of dying from prostate cancer in 10 years is 5 in 1,000. The chance of a screened man dying is either 4 or 5 in 1,000, so the net benefit is between 0 and 1 man per 1,000 men screened.

     7) NO study showed a reduction in all cause mortality(!) The one study that showed a slight decrease in mortality from prostate cancer showed an increase in deaths from other causes, usually cardiovascular, so there was no net extension of lifespan that resulted from PSA screening.

     8) When the PSA cutoff is between 2.5 and 4.0, the false positive PSA rate is 80%.

     9) Overdiagnosis makes screening appear to save lives when it does not do so.

     10) The median follow-up time was 10 years.

     11) Because of lead-time bias, patients diagnosed earlier by asymptomatic screening appear to live longer with cancer. This is a problem with all screening methodologies for all fatal diseases.

     12) The American Cancer Society and the American Urological Society both recommend informed decision making rather than routine PSA screening.

     13) Radiotherapy and surgery both cause urinary incontinence in 20% of those so treated and erectile dysfunction in 30% (on the average).

     14) Radiotherapy can also cause chronic diarrhea and/or colonic irritability.

     15) "We need to practice medicine on the basis of evidence and not on the basis of faith."

     And again, despite all recommendations and readings, there is no substitute for a full discussion with your family doctor of your likes, your dislikes, your wishes, your fears, and your desires. You should never feel pressured into making a decision about your health with which you do not feel comfortable. It is your body and your mind, and you have to live comfortably with them both. And once you make a decision, please don't second-guess yourself. (You  have enough well-meaning friends who will do that for you.)



  1. I understand the puzzlement of patients. They were told that early detection was the road to cure, and this does not seem to apply to using the PSA to detect prostate cancer. This cancer just has different properties.

  2. One final note: In a study just published in the (July 19) New England Journal of Medicine, with a median follow-up of 10 years of patients with localized prostate cancer, there was no difference in the death rate from prostate cancer between watchful waiting and radical prostate cancer surgery.

  3. Final final note: There was a 44% reduction in prostate cancer mortality between 1993 and 2009.

  4. Expect the following with less PSA screening: more cases of advanced prostate, more deaths from prostate cancer, and screams from the press and other interested parties that we are killing men by not getting PSA testing done. The truth of it is that our early detection and reduction of deaths from PSA was offset by the morbidity and mortality from treatment, which won't show up in statistics. This is harder to describe and makes poorer headlines, so the PSA lovers will have the advantage. It doesn't hurt their cause that this can be made into a political issue under the incorrect label of "something bad caused by Obamacare."

    I think that one of two things needs to be done to make PSA more useful: either there needs to be a simple test with high specificity for the more aggressive form of prostate cancer (that men die from, not with), or there needs to be a genetic screening test that gives us a subgroup of men who are at high risk for aggressive prostate cancer. We can do PSA on the high-risk men and have more faith that the result is significant.