Sherman Frankel, Dept. of Physics, University of Pennsylvania
May, 1996
The PSA score is widely used as a measure of onset of cancerous activity in the prostate gland. Strong fluctuations in the score are often a cause of alarm on the part of the patient and invasive action on the part of urologists to determine whether cancerous activity is present and a real danger.
Recently the importance of reporting and understanding correlations in the improvements in variables for finasteride (1) and in the side effects for terozosin (2) were emphasized. This comment deals with correlations between PSA scores and current illnesses at the time the blood sample is drawn. While the data presented are for a single case, the large magnitude of the effects and the mathematical description of the progressive change in the PSA may make this report more than anecdotal.
Recently, my physician, in treating me for a stomache ailment, noticed I had not had my periodic PSA test and suggested I get a reading. The score was 43.5 whereas my previous steady average over several years was 4-5! Yet, fluctuations of only 15% are claimed to be significant (3), and the recommendation for 3 measurements (4), and a suggested retest interval (5) have been discussed.
I had once noted a jump from 4 to 9 immediately after a urological exam and had observed it to return to 5 a week later. But this new jump was huge by any standards. Accordingly I decided to follow the process over a longer period; the first measurement might simply have been in error. The enclosed figure shows a plot of the PSA score vs days on a logarithmic plot. I chose that display because the rate of fall, if it occurred, might be proportional to the score. As we can see from the linear falloff on the logarithmic plot, this is the case, the half life for ``recovery'' being about 5 days, the fall appearing to be exponential, varying as .
It seems to me that this is a remarkable observation. I had no apparent stressing of the prostate and had no prostate examination in the preceding months. I was relieved to find that the recovery was complete. However, as the figure shows, I chose to verify that equilibrium had been reached by taking another score at a later period, to assure that there was indeed a reasonably reproducible base line.
This large fluctuation and its rate of recovery suggests that there may be a large and not fully appreciated correlation between PSA fluctuations and other coexisting illnesses as well. Such a study seems advisable and should not be difficult to make. Older patients with BPS are likely to take PSA tests when visiting physicians for other illnesses, perhaps suggested by a physician who notes a long period has elapsed between tests. These records should be available in any hospital data base and one might easily be able to search for such correlations. Knowing of their existence would certainly affect how urologists respond to a sudden fluctuation. By being alerted and looking for the presence of a fall-off under such circumstances, some invasive scenarios could be avoided. Further, the presence of correlations, if further study supports our supposition, would add to our medical understanding.
psacorr; printed June 16, 1997
1. "Analyzing Finasteride Data", S. Frankel, Neurourology and Urodynamics, 14:619-624 (1995)
2. "Assessing Hytrin Side-Effects", S. Frankel, Mar. 5, 1996, I/P 9605c, to be published
3. Dov Kadmon et al., Journal of Urology, 155 1665 May 1996
4. H. B. Carter et al. Sem. Oncol. 21 554 (1994)
5. H. B. Carter et al. Urology 45 591 (1995)