Baseline PSA and prostate cancer-specific mortality


It is well understood that a diagnosis of prostate cancer does not necessarily imply that the patient will ever have progressive disease — let alone that he will die of prostate cancer. Indeed, the number of men in America who die of prostate cancer today is believed to be significantly less than three for every 100 men diagnosed.

However, a recent analysis of long-term data from Duke Medical Center does suggest that the baseline PSA level of a patient subsequently diagnosed with prostate cancer (i.e., his PSA level at the time of his first PSA test) may be highly relevant to his personal risk of prostate cancer-specific mortality.

The new paper by Tang et al. (in concert with information in a media release from Duke Medical Center) reports that the Duke Prostate Center database was used to identify men who had their PSA level measured over the past 20 years. The available data were used to show the following:

  • The database included 4,568 men diagnosed with prostate cancer after a baseline PSA measurement.
  • The median age of these patients was 65 years at the time of initial PSA testing.
  • The median baseline PSA was 4.5 ng/ml.
  • The average (mean) follow-up period was > 9  years.
  • 3.5 percent of the men died from prostate cancer during the study period; > 20 percent died from other causes.
  • A baseline PSA of ≥ 10 ng/ml was 11 times more likely to be associated with prostate cancer-specific mortality than a baseline PSA level of  < 2.5 ng/ml.
  • A baseline PSA level between 4.0 and 9.9 ng/ml was three times more likely to be associated with prostate cancer-specific mortality than a baseline PSA level of  < 2.5 ng/ml.
  • An advanced age at baseline PSA and African American race were associated with a higher death rate from prostate cancer and death from all causes.

Now it is worth emphasizing a specific point related to this study. It only applies to men who are actually diagnosed with prostate cancer.

This study does not address information about the PSA levels of all the other men who may have been seen at Duke in the same 20-year timeframe and who had a PSA level that may have been elevated as a consequence of prostatitis, or benign prostatic hyperplasia, or trauma, or any other possible stimulus. Thus, we cannot simply say that a baseline PSA level of ≥ 4 ng/ml places a man at heightened risk for prostate cancer-specific mortality. We have to be clear that what this study tells us is that a baseline PSA of ≥ 4 ng/ml in a man who is subsequently diagnosed with prostate cancer increases that man’s risk for prostate cancer-specific mortality by comparison with a baseline PSA of < 2.5 ng/ml.

The critical learning issue here seems to be that early detection of the cancer is key to the potential to minimize the risk of prostate cancer-specific mortality. It should not simply be assumed that early detection of a PSA of ≥ 4 ng/ml in an inevitable signal of heightened risk for prostate cancer-specific mortality.

It is also worth bearing in mind that the data accumulated in this study do go back 20 years. In those 20 years a lot has changed, the average baseline PSA level of 4,500+ men diagnosed today would probably be significantly lower, for example, than the median baseline PSA of the men in this study, because 20 years ago most men didn’t get baseline PSAs in their early 50s (or younger). Indeed, in this study the median age at the time of the baseline PSA test was 65! So time may have an important impact on how to interpret these data with accuracy.

6 Responses

  1. I just keep reading Dan Zenka’s story. If there’s anyone in the world who should have had access to the resources and knowledge to be properly screened for early diagnosis, it would be Dan. And yet he was fundamentally failed by the lab/biopsy cycle of contradiction and misinformation.

    Is it better to start from the perspective of dismissing the potential seriousness of a situation and run the risk of engaging too late or is it better to treat a situation as deadly serious — act with due diligence and haste and the run the risk of “over” treating and overspending resources? Seems that choosing the former has cost Dan a LOT.

  2. Without knowing whether Dan made a specific decision not to have a biopsy when his PSA was only 3.0 a year earlier, I don’t think we are in a position to know whether that decision cost Dan anything. And if he did make an informed decision not to have a biopsy in 2009, then he must have known that he was taking a risk.

    As I am quite sure Dan would agree, this is why quibbling over the details of what PSA data mean is unhelpful by comparison with the need to find a much better test with a high degree of specificity for aggressive disease.

  3. And as far as I can tell from your postings of the last several months there’s NOTHING on the horizon even close to qualifying as “a much better test with a high degree of specificity for aggressive disease.”

    How many men’s lives have to be ruined before something reliable and useful makes it to market?

  4. Sadly, nothing like as many as the number of men whose lives are ended each year by allowing people who are drunk or incompetent to drive cars … and yet social wisdom says you can get a driving license after demonstrating only the most rudimentary level of competence. We live in a sadly imperfect world. Should we take away the driving licenses of 30% of Americans?

  5. Hi Mike,

    My good friend. Please forgive me but I once again I lay protest to the funny math used to calculate prostate cancer mortality. This year it is predicted that 28,000+ will die of prostate cancer. 200,000+ will be diagnosed. That’s 14 of every 100 men diagnosed that will pass due to prostate cancer. But before you say “but not all those men were diagnosed this year” stuff I understand that if what you are saying is 3 of every 100 men alive with prostate cancer will die this year you can understand my protest is based on what you wrote.

    But that is funny math that aggregates over years of survivors compared to the number of men that will die this year. To what value THAT number is ~ is a complete mystery to me. But since the same approximate number of 14% deaths/diagnosis ratio has been consistent over the last decade and longer, I can’t buy into your statement that less than 3 of every 100 diagnosed will die of the disease. Sooner rather than later a generation will pass in the PSA testing era and that the same percentage of 14 in 100 will show that if you add up the number of men diagnosed, and the number of men deceased from prostate cancer, over all years, it will be much higher than 3% death to diagnosis ratio.

    Now, back to finding a good place for YP.

  6. Tony:

    Don’t shoot the messenger. Talk to the statisticians.

    We know that the model the American Cancer Society uses to predict numbers of patients and deaths each year isn’t accurate. They admit this themselves — and that’s just for starters! If you look at the SEER data (as I showed you when we were in LV), the actual mortality rate has been dropping for years.

    My personal suspicion is that almost all of the “numbers” are manipulated by people trying to make specific points. There are three things I feel pretty comfortable about: (1) It ain’t actually as high as 14% (relative to the number of patients diagnosed on an annual basis). (2) It ain’t zero relative to the number of men living after a diagnosis (now believed to be ~ 2.2 million)! (3) We will never completely eliminate the possibility of death from prostate cancer — although we may be able to get pretty close if we continue to make progress at the current pace.

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google 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: