This edition of the Cleveland Clinic Journal of Medicine includes a timely update on prostate cancer screening and prevention by a leading international expert, Dr. Eric Klein.1 At long last, 2009 brought the publication of two large prostate cancer screening trials.2,3 Randomized controlled trials had been needed to discover whether screening had a benefit.
Now that we have the data, was it worth the wait? Do we know the answer? Should our male patients, our male loved ones, and those of us who are men have prostate-specific antigen (PSA) tests?
DOES EARLIER DIAGNOSIS HELP OR HARM?
Over the past 20 years, PSA screening and other developments have transformed the presentation of prostate cancer in regions where PSA testing is common. The incidence of prostate cancer that was metastatic at the time of diagnosis fell by 56% between 1985 and 1995.4 The proportion of cancers that were localized in the mid-1980s was 58%, compared with 80% now, while only 4% now have metastases at diagnosis.5
This early detection had a predictable effect on 5-year relative survival, which increased from 69% in the mid-1970s and 84% in the late 1980s to 99.9% in the early 21st century.5 Prostate cancer now has the highest 5-year relative survival of any cancer except non-melanoma skin cancer.
This doesn’t mean that prostate cancer doesn’t kill men, but only that it almost always takes longer than 5 years from diagnosis. More than 27,000 Americans die of prostate cancer annually—lung cancer is the only malignancy that kills more men. Nonetheless, that 27,000 is a small fraction of the 192,000 men diagnosed with prostate cancer each year. And it is worth keeping in mind that autopsy studies show that most men have cancer in their prostates by the time they reach age 70, while the Prostate Cancer Prevention Trial reported that 24% of men at least 55 years old have prostate cancer detectable by biopsy, including 15% of men who have a serum PSA less than 4.0 ng/mL and a normal digital rectal examination.6,7
Prostate cancer is thus highly prevalent, usually indolent, but sometimes deadly. Overtreatment of indolent disease and ineffective treatment of aggressive disease continue to represent major challenges.
As prostate cancer survival has lengthened, the prostate cancer death rate has declined, although to a lesser extent. The death rate from prostate cancer per 100,000 US males was 31 in 1975, climbed to 39 in 1990, and then declined to 25 in 2005, a 19% reduction over 30 years. Viewed differently, the lifetime risk of being diagnosed with prostate cancer increased from 13% in 1990 to 16% in 2006, while the risk of dying from it declined from 3.2% to 2.8%.5
This reduction in death rate was interpreted by some as evidence that PSA screening is effective, but it was impossible to control for confounding variables such as improvements in treatment. It was clear that PSA testing provided earlier diagnosis and hence longer survival from the time of diagnosis, but it was not clear whether it resulted in men living longer. Given the numerous kinds of serious harm that can follow from a diagnosis of prostate cancer in the form of anxiety, treatment side effects, and medical expenses, early diagnosis could easily represent a net harm.
THE EUROPEAN PROSTATE CANCER SCREENING TRIAL
To address the question of whether prostate cancer screening with PSA testing lowers a man’s risk of dying of prostate cancer, Europe and the United States each initiated randomized controlled trials.
The European study2 randomized 162,000 men, age 55 to 69 years, to one of two groups. One group was offered PSA screening, the other was not. In those screened, PSA testing was repeated once every 4 years on average. Most centers participating in the trial used a PSA above 3.0 ng/mL as the threshold for biopsy. In the screening group, 82% of the men had at least one PSA test, 16% of all PSA tests were positive, and 86% of men who had an elevated PSA value underwent a biopsy. Of those undergoing biopsy for an elevated PSA, 76% had benign results, which shows that PSA as a test for cancer has a high false-positive rate.
As expected, screening increased the rate of prostate cancer detection. The rate was 70% higher in the screening group: 8.2% of men in the screening group were diagnosed, compared with 4.8% in the control group. Men undergoing screening were more likely to have localized disease and 41% less likely to have metastatic disease. The increased number of cancers detected by screening were predominantly less-aggressive tumors: the incidence of low- and intermediate-grade cancers (Gleason score 2 to 6) was 4.8% in the screened group vs 1.7% in the control group. Screened men had a lower proportion (28% vs 45%) but a higher incidence (1.9% vs 1.4%) of high-grade cancers (Gleason score 7 or higher). It is this tendency of screening to preferentially detect indolent cancers that results in length-time bias.
So did PSA testing lower the risk of death from prostate cancer? In the European trial it did, and by 20% (95% confidence interval 5%–33%, P = .01). This study thus provided level-1 evidence that PSA testing to screen for prostate cancer reduces prostate cancer mortality rates.
However, more than 1,400 men needed to be screened and 48 needed to be treated for each death prevented. Moreover, because fewer than 3% of men die of prostate cancer, lowering the risk of death from prostate cancer does not result in an appreciable effect on all-cause mortality or on life expectancy. We cannot say that men live longer as a result of prostate cancer screening—only that they are less likely to die of prostate cancer.