PSA SCREENING: THE DEBATE GOES ON

Screening reduced prostate cancer–specific mortality slightly in one study, and not at all in the other.

The idea that screening reduces prostate cancer mortality has largely been a leap of faith. Now, finally, we have a first look at outcomes from two randomized screening trials -- and the results are ambiguous.

In the PLCO trial from the U.S., 77,000 men (age range, 55--74) were randomized to screening (annual prostate-specific antigen [PSA] testing for 6 years plus annual digital rectal examination [DRE] for 4 years) or no screening. Men who had PSA levels >4.0 ng/mL or suspicious DREs were advised to seek further evaluation, but subsequent diagnostic and therapeutic interventions were not standardized. Through year 10, there were 92 prostate cancer deaths in the screening group and 82 in the control group, a nonsignificant difference. Overall mortality also was similar between groups. PSA testing outside the trial was considerable: The authors note that among controls, "the rate of PSA testing was 40% in the first year and 52% in the sixth year," but they do not clearly state whether these rates are cumulative or year-by-year.

In the ERSPC trial from Europe, 182,000 men (age range, 50--74) in seven countries were randomized to PSA screening or no screening. The seven countries varied considerably in PSA screening intervals, PSA cutoffs for further evaluation, and inclusion of DRE in initial screening. During an average follow-up of 9 years, there were seven fewer prostate cancer deaths per 10,000 screened men compared with controls (adjusted P=0.04). To prevent one prostate cancer death, 1410 men had to be screened, and 48 additional cases of prostate cancer had to be diagnosed and treated. All-cause mortality did not differ in the two groups. The rate of screening among controls was not clearly stated.

Comment

In the U.S. study, prostate cancer--specific mortality was unaffected by screening; in the European study, a reduction in mortality barely reached statistical significance. Both studies had important limitations. For example, U.S. controls underwent considerable screening outside the trial; in the European trial, pooling the outcomes of seven somewhat disparate protocols is problematic. Screening enthusiasts will likely embrace the European trial and downplay the U.S. trial; the reverse will be true for skeptics. Additional follow-up, more information on the burdens and downstream complications of screening, and cost-effectiveness analyses are presumably forthcoming from both trials. In the meantime, what should clinicians tell patients? The editorialist's conclusion is a reasonable starting point: "Serial PSA screening has at best a modest effect on prostate-cancer mortality during the first decade of follow-up. This benefit comes at the cost of substantial overdiagnosis and overtreatment."

In my view, an important next step is for a neutral body such as the U.S. Preventive Services Task Force to update its analysis of prostate cancer screening, given these new data (JW Aug 26 2008). Healthcare policy makers also need to ask boldly whether the PSA screening juggernaut -- with all the time, energy, and resources consumed by screening and its sequelae -- is appropriate in an era of unsustainable growth in healthcare spending.