1. Home>
2. Specialties>
3. General Medicine>
4. Top Story

PSA Screening Lacks Value

Proponents of using serum prostate-specific antigen (PSA) assays for prostate cancer screening have assumed that its value and accuracy would become clearer as its use was refined through new research. Several studies from 2004 detract from that enthusiasm.

Most provocative was a study in which 2950 men (age, 62) with normal rectal exams and PSA levels 4.0 ng/mL underwent prostate biopsies. Cancer was found in 15% of patients, ranging from 7% in men with PSA levels 0.5 ng/mL to 27% in men with levels of 3.1-4.0 ng/mL (Journal Watch Jun 8 2004). For some experts, these results suggested that PSA threshold levels for recommending biopsies should be lowered. However, lack of outcome data on cancers that are detected by such biopsies make this recommendation problematic.

The findings of a study on the psychological effects of performing biopsies in men with elevated PSA levels suggested that we should be cautious in endorsing an expanded biopsy protocol. When 167 men who underwent biopsies were compared with 233 men who had normal PSA levels (<2.5 ng/mL) and did not undergo biopsies, the men who underwent biopsies were more likely to worry about developing prostate cancer (40% vs. 8%; Journal Watch Dec 17 2004).

The controversial nature of this issue perhaps is conveyed best by a recent article entitled The PSA era in the United States is over for prostate cancer (Journal Watch Oct 29 2004). Researchers reviewed specimens from 1317 radical prostatectomies for localized prostate cancer, obtained from 1983 through 2003. In the early years, when nearly all cases were detected by digital examination, preoperative PSA levels correlated with tumor aggressiveness (i.e., tumor volume, Gleason score). But in the later years, when nearly all cases were detected by PSA screening, preoperative PSA levels were highly correlated only with total prostate weight. This finding implies that mildly elevated screening PSA levels in later years mainly reflected benign hyperplasia, and that the cancers were incidental findings on biopsy. According to this reasoning, lowering the PSA threshold for biopsy would simply worsen the problem of overdiagnosis of small tumors with benign characteristics.

The annual rate by which PSA rises (i.e., PSA velocity) is one way to improve the prognostic accuracy of PSA screening. In a study of 1095 men with localized cancer (most detected with PSA screening), PSA velocity before diagnoses of prostate cancer was associated with eventual outcomes after prostatectomy: 5-year mortality rates were 9% for men with annual PSA velocities >2.0 ng/mL, and 0.3% for men with lower velocities (Journal Watch Jul 13 2004).

While we await the results of randomized trials to determine whether PSA screening reduces mortality, many clinical organizations recommend that physicians have discussions with individual patients about risks and benefits of PSA screening. Those discussions have become longer and more complicated because of the PSA research reported in 2004.

— Thomas L. Schwenk, MD

Published in Journal Watch General Medicine December 31, 2004