PSA and Prostate Cancer Screening

Prostate cancer is an important disease in men. It is estimated that in 2010 there were expected to be 218,000 new prostate cancer diagnoses and about 32,000 prostate cancer deaths, making it the second most common cause of cancer death in men after lung cancer.

Virtually all men are now aware of this disease, and many have heard of the blood test PSA (prostate specific antigen) used to screen for prostate cancer. Not so commonly known are the uncertainties regarding the value of this test.

Medical organizations around the world have published widely differing recommendations regarding screening for prostate cancer and the use of PSA. These range from the American Urological Association recommendation that this test be discussed and potentially offered to all men over 40, to the Canadian recommendation that it be offered to no one.

Often, disagreements about issues such as these arise from a lack of information, a lack of well done clinical trials. In such trials, volunteers agree to be randomly assigned to one or more treatments or testing strategies and are carefully followed to see which treatments, or testing strategies, lead to better outcomes.

Surprisingly, PSA screening has been subjected to two very large trials lasting 5-10 years and involving over 200,000 men. So why is the value of this test uncertain?

There are multiple reasons, including the complex and unusual nature of prostate cancer, the high rate of abnormal PSA levels in patients without cancer, the uncertainties about the effectiveness of current prostate cancer treatments, and the high side effect rates of these treatments.

For an American male, the lifetime risk of being diagnosed with prostate cancer is 16%, but the risk of dying of prostate cancer is only 2.9%. Even higher is the percentage of men who get prostate cancer but it is never diagnosed. Studies in which all men over a certain age had prostate biopsies, or evaluations were performed on the prostates of all men who died from any cause, show that prostate cancer is present in almost 50% of men by age 80, and nearly 100% of men in their 90s. These data suggest that prostate cancer often grows so slowly that most men die of other causes before the disease becomes clinically advanced, and that even if it is aggressive enough to be diagnosed clinically, the odds of dieing from the disease are less than one in five.

Prostate cancer survival is related to many factors, especially the size and spread of the tumor at the time of diagnosis. The 5-year survival among men with cancer confined to the prostate (localized) or with just local spread is 100 percent, compared with 31.9 percent among those diagnosed with distant spread of the tumor. This, and other data suggest that men with distant spread of the tumor can not generally be cured. Thus, a screening test or tests that could identify asymptomatic men with aggressive tumors that have not yet spread to other areas might be expected to substantially reduce prostate cancer death.

In the United States PSA at or above 4.0 (ng/mL) is considered to indicate potential prostate cancer. In Europe a value at or above 3.0 is used. In large studies, the actual odds of having prostate cancer for PSA levels between 4.0 to 10.0 is approximately 25 percent, and is 42 to 64 percent for PSA levels >10. However, nearly 75 percent of cancers detected with PSA values between 4.0 to 10.0 are confined to the prostate and are potentially curable. The proportion of organ-confined cancers drops to less than 50 percent for PSA values above 10.0. Thus, detecting the curable cancers in men with PSA levels less than 10.0 presents a diagnostic challenge because about ¾ of biopsies in this group are unnecessary. Various modifications of PSA testing have been tried, but only two have shown to help. First, because PSA varies in men without cancer, some experts recommend first repeating an abnormal PSA several weeks later if it is below 7.0. Before repeating a PSA between 4.0 and 6.9, patients should be asked to refrain from sexual activity and bike riding for at least 48 hours, because these can falsely elevate PSA, and, if there is evidence of prostate inflammation, complete a course of antibiotics. Second, when the test is repeated, measuring free and total PSA, a slightly more expensive variant of PSA testing, occasionally adds enough diagnostic accuracy to make a clinical difference. Other tests such as PSA velocity and PSA density have not been shown to be effective.

Does treatment of prostate cancer detected using PSA screening work (does it decrease death from prostate cancer)? The main treatments are radical prostatectomy (removing the prostate and the immediately adjacent tissues surgically), external beam radiation of the prostate (destroying the prostate and the surrounding tissues with radiation), or brachytherapy (putting small particles of radioactive material in the prostate to destroy it). Another option is called active surveillance, which means not immediately treating a cancer but following PSA tests, rectal exams, and repeating biopsies to determine whether aggressive treatment is indicated because the cancer is progressing.

A study of 875 men undergoing radical prostatectomy found only a limited association between preoperative PSA levels of 2 to 9 and cure rates. The disease-free survival curves did not significantly diverge until the preoperative PSA levels reached 7, suggesting that diagnosing cancers at a lower PSA level may be unnecessary, and that radical prostatectomy may not improve survival at a PSA under 7 (possibly because a very high percentage of patients would have survived without any treatment).

The largest and longest direct trial of PSA screening to date, involving 182,160 men between the ages of 50 and 74 followed for an average of 9 years, showed a 20% decrease in death from prostate cancer in men screened with PSA once every 4 years. In this trial 1410 men needed to be screened to prevent one prostate cancer death over nine years, and 48 additional patients were diagnosed with prostate cancer to prevent one prostate cancer death.

What harms are done by testing people using PSA?

As stated above, in the largest direct trial, 1410 men needed to be screened with PSA once every 4 years, approximately 150 men had to be biopsied, 48 additional men were diagnosed with prostate cancer, and approximately 36 men treated for prostate cancer to prevent one prostate cancer death over nine years. An observational trial looking at all prostate cancer deaths in the US from 1986 through 2005 estimated that if 100% of the decrease in prostate cancer deaths which have occurred in the first nine-years after the introduction of PSA testing were due only to PSA testing, and not to any other advances in the diagnosis or treatment of the disease (which is the maximum possible benefit from PSA testing), then approximately 23 men had to be diagnosed and 18 men treated for prostate cancer as a result of PSA testing to prevent one death.

According to these two studies, 2800 PSA tests, 102 biopsies in patients without cancer, 5-12 men newly diagnosed with cancer who are not directly treated but who now have a diagnosis of cancer, and 18-36 men newly diagnosed with cancer who are treated without change in their survival, are necessary to prevent one death over 9 years. The harm of PSA testing is monetary and is about $56,000. The harm of the biopsies is monetary, about $55,000, and pain, which is generally not severe. The harm to the 9 men who are not treated (and presumably undergoing active surveillance), is the same as PSA and biopsy, plus the cost and pain of ongoing surveillance, and the psychological and insurance burdens of carrying a diagnosis of cancer. The harm to the 27 men who are treated without change in survival is the same as the men who are not treated plus the cost and harm of treatment. Cost of treatment is about $500,000. Harms include 0.5% risk of death, 35% risk of erectile dysfunction, 15% risk of urinary incontinence, and 10% risk of bowel dysfunction. This indicates that a minimum of 16 men will have either permanent erectile dysfunction, urinary incontinence, or bowel dysfunction for every man whose life is saved.

The above mentioned study of US prostate cancer screening using PSA from 1987 to 2005 estimated that in these 8 years 1.3 million additional men were diagnosed with prostate cancer as a result of PSA screening, of whom approximately 1 million were treated.

A recent statement from the American College of Physicians summarizes the above information as follows: Although screening for prostate cancer with PSA can reduce death from prostate cancer, the absolute reduction is very small. Given limitations in the design and reporting of the randomized trials, there remain important concerns about whether the benefits of screening outweigh the potential harms to quality of life, including the substantial risks for overdiagnosis and treatment complications. Men who are willing to accept a substantial risk of side effects associated with treatment in return for a small improvement in survival might reasonably choose to be screened. Men who are at increased risk of prostate cancer because of race or family history may be more likely to benefit from screening.