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) determinations, as part of an early detection program. Indeed, many prefer to rely on the PSA level alone, rather than have the more intrusive digital rectal exam. There's no doubt that the use of PSA testing has resulted in an increase in the number of men having prostate biopsies and subsequently being diagnosed with cancer, with a shift towards detection at earlier stages.

Along with earlier detection, there has also been an increase in the numbers of radical prostatectomies done for small cancers that might, indeed, be clinically insignificant; one estimate puts the frequency of this at 26%. It's been asked if widespread PSA testing is directly responsible for unnecessary surgery. The specificity of the test has been questioned when results fall in the moderately increased range (4 to 10 ng/mL). To help clarify this, US investigators have studied the variability of PSA results over time. Previous attempts have been made to do this, but the numbers of subjects were small and the period of observation was relatively short.

Method

The scientists used blood samples taken from men participating in a colon polyp prevention study designed to evaluate the effect of a diet low in fat and high in fiber, fruits, and vegetables. Five consecutive blood samples from 972 men, with a mean age of 62, were obtained over a 4-year period. Total PSA levels were determined in all samples, while free PSA was measured as well if the total PSA was between 4 and 10 ng/mL.

Results

Mean total PSA levels increased with age, as shown in the table below:

Age group	No. of Men	Mean (95% CI)
<50	102	0.8 (0.7-1.0)
50-59	289	1.5 (1.3-1.7)
60-69	371	2.3 (2.0-2.5)
70-79	201	2.8 (2.4-3.2)
>79	9	4.4 (0 -10.2)

There are several criteria for determining the need for a prostate biopsy: PSA level >4.0 ng/mL; PSA level >2.5 ng/mL; above the age-specific PSA level; free-to-total ratio of <0.25 ng/mL; and PSA velocity >0.75 ng/mL per year. Using these criteria, it was found that between 15% and 37% of the men would have qualified for a biopsy, depending on the criterion applied; 15% using PSA velocity and 37% using PSA >2.5 ng/mL would have qualified.

Anyone who developed prostate cancer during the 4-year period was excluded from subsequent analyses. These comprised an assessment of what proportion of abnormal PSA levels returned to normal the year after it had been elevated, at any later visit, and had normal levels for two consecutive tests. The findings are given below:

PSA Criterion	% Normal at Next Year Visit	% Normal at Any Later Visit	% Normal on 2 Consecutive Visits
>4.0 ng/mL	30%	44%	80%
>2.5 ng/mL	26%	40%	65%
Above age-specific level	37%	55%	83%
Free: total ratio <0.25 ng/mL	35%	53%	74%

Comment

In this study, prostate biopsy would have been recommended in 21% of the men, based on a single PSA result above 4.0 ng/mL over a 4-year period. This corresponds fairly well to the findings in large-scale prostate cancer screening programs - 10% to 15% in isolated sampling.

What's important about the results of this analysis is the finding that nearly half of men with one abnormal PSA level subsequently had a normal level. In other words, variations in PSA levels may lead to many false-positive test results, and possibly unnecessary biopsies.

It's known that prostatitis can cause elevated PSA levels. A practical approach, therefore, in the absence of any clinical indications of carcinoma, would be to have the subject take a course of an antibiotic and an anti-inflammatory, and return in 4-6 weeks for a second PSA determination. This policy could, theoretically, allow growth and spread of a malignant tumor during the 6-week waiting period. However, this type of tumor has a slow natural history, as shown by the success of 'watchful waiting' programs, so that this risk is not substantial.

This study has performed a useful service in helping to allay some of the anxiety associated with an elevated PSA result, as well as supporting the concept of a measured approach to the problem that should have economic as well as 'good medicine' benefits.