Introduction

Cancer of the prostate is the second most common cause of cancer deaths in men, after lung cancer. Not surprisingly, more and more men are having regular serum prostate-specific antigen (PSA) tests, as part of early detection programs. The use of PSA testing has resulted in more men having prostate biopsies¹ and subsequently being diagnosed with cancer. This means that prostate cancer is being diagnosed more often, and in earlier stages, than before. And, perhaps, surgery to remove the prostate is being done for very small cancers that might, in fact, not need to be removed.

Is it possible that widespread PSA testing is directly responsible for unnecessary surgery? It seems that PSA test results can vary quite a lot from time to time, and that an abnormal (raised) result may not be a 'true' indication of prostate disease. This problem has been studied in analyses carried out on blood samples from men enrolled in a screening program for something entirely unrelated (colon polyps).

What was done

The colon polyp prevention study looked at the protective effect of a diet low in fat and high in fiber, fruits, and vegetables. Five consecutive blood samples were obtained from 972 men (average age 62) over a 4-year period, and the PSA levels measured. Samples from anyone who had prostate cancer, or who developed it during the 4-year period, were excluded from the analyses. (It should be noted that the blood samples were analyzed for PSA long after the colon polyp study had finished, so that there was no question of possibly 'missing the diagnosis' of prostate cancer.)

What was found

The average PSA levels increased with age, as shown in the table below:

Age group	No. of Men	Average PSA (ng/mL)
<50	102	0.8
50-59	289	1.5
60-69	371	2.3
70-79	201	2.8
>79	9	4.4

Although the full analysis used several different measures for labeling a result "abnormal" (and therefore calling for a prostate biopsy), in this summary we'll just look at the results they got using just one; levels above 4.0 ng/mL total PSA.

Using this standard, it was found that 21% of the men - one in five - would have qualified for a biopsy, and some of these might have had findings on microscopy that would lead to treatment for prostate cancer (surgery, radiation seeds, or the like).

Further analyses looked at what percentage of men with abnormal PSA levels had a normal result the next year, or at any subsequent testing. It was found that 30% - about a third - of the men with a raised PSA had a normal value a year later, and as many as 44% - almost half - had a normal result at some time after their abnormal result.

What does this mean?

The variations in PSA levels of individual men from year to year show that one shouldn't place too much reliance on a single raised reading (unless there are other suspicious signs of prostate cancer, such as nodules found by the physician on digital rectal examination). Indeed, if one had relied on single readings, 21% of the men in the colon polyp study would have qualified for a prostate biopsy.

Prostate biopsy is not an entirely trivial procedure; infections have been reported in about 5% of cases, with blood in the urine afterwards in 3%. It would be cheaper, and better medical practice, if unnecessary biopsies could be avoided.

The important results of this study show that nearly half the men with one abnormal PSA level subsequently have a normal value. In other words, variations in PSA levels may lead to many false-positive test results.

It's known that prostatitis (inflammation of the prostate) can also cause raised PSA levels. A practical approach, therefore, would be to have the patient take a course of an antibiotic and an anti-inflammatory drug, and return in 4-6 weeks for a second PSA determination. Theoretically, this policy could allow the growth and spread of a malignant tumor during the 6-week interval. However, this type of tumor grows very slowly, as shown by the success of 'watchful waiting' programs, so that this is not a genuine risk.

This study is helpful, as it can lessen some of the anxiety associated with a raised PSA result; it also shows that a more restrained approach - i.e. a second PSA test after an interval of 4-6 weeks, with or without an antibiotic - may save unnecessary procedures that are not always very comfortable.