Swedish study: PSA screening may cut prostate cancer mortality in half

Population screening with the PSA test in men ages 50 to 69 years cut prostate cancer mortality nearly in half during a 14-year follow-up, according to data from a Swedish study published July 1 in the online edition of Lancet Oncology.

"However, the risk of over-diagnosis is substantial and the number needed to treat is at least as high as in breast-cancer screening programs," wrote the authors, led by Jonas Hugosson, MD, of the University of Gotenburg. "The benefit of prostate cancer screening compares favourably to other cancer screening programs."

The Gotenburg study began in 1994, enrolling 20,000 men who lived in the city and were older than 50 years (median age, 56 years). Men were randomized to either a screening group invited for PSA testing every 2 years or to a control group that was not invited. Invitations in the screening group were discontinued at a median age of 69 years.

This study shows that a PSA-based screening initiative is approved of by men aged 50 years or older, as 76% of men attended at least once, according to the study authors. During the median 14-year follow-up, prostate cancer was diagnosed in 12.7% of men in the screening group and 8.2% in the control group (hazard ratio, 1.64; p

The researchers reported that there were 44 deaths from prostate cancer in the screening group and 78 in the control group. The screening group saw a rate ratio of 0.56 of dying from prostate cancer, compared with the control group (p=.002). Absolute cumulative risk reduction of death from prostate cancer at 14 years was 0.40%, compared with 0.90% in the control group and 0.50% in the screening group.

In an accompanying editorial, David Neal MD, of the University of Cambridge, United Kingdom, points out that the study was small (20,000 men), and that the finding is being generated from "a country with low levels of opportunistic PSA testing," in direct contrast to the situation in the United States, where widespread PSA testing is common. As a result, the results may not be generalizable to the United States.