Some 16 years after randomized trials of PSA screening commenced in Europe and the US, many people anticipated that they would provide “the definitive answer” to this thorny issue. Both trials published their preliminary findings in the March 26, 2009 issue of the New England Journal of Medicine, and the results are apparently conflicting. This article helps interpret the studies (the full texts are available on the NEJM website at http://content.nejm.org).
The PLCO screening trial
The American Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial was the smaller of the two studies; approximately 77,000 men between the ages of 55 and 74 years were randomized (i.e. selected at random) to participate in a group that had a PSA test and DRE (digital rectal examination) every year for six years, or to be in the “control” group. All the men were followed for an average of 11 years. In the screening group, if the PSA was higher than 4 ng/mL or the DRE was abnormal, the family doctor was informed and the patient and doctor decided on the next step. For some, this meant repeating the PSA test; if it fell below 4 ng/mL, no action was taken. The result was that only two-thirds of the men had a biopsy. Men in the control group were not targeted for PSA testing, but because PSA testing is popular in the US, about half of these men had a PSA test either before or during the trial anyway (known as study contamination).
For the men in the PSA screening group, the trial produced a 20% increase in the diagnosis rate of cancer. However, the stage and grade of cancer diagnosed was similar in both the screening and control groups: 95% of the men had localized cancer, even though you would have thought there would be significantly more advanced cancers in the control group (this similarity was likely a result of contamination). So it’s not surprising that there was no reduction in the rate of prostate cancer deaths due to screening — even after 11 years, one would expect only men with high-risk or metastatic cancer to have succumbed to the disease. On the other hand, a sub-group analysis showed a reduced death rate of 25% for men who had at least two PSA tests versus those who had none (whether they were assigned to be tested or not). This may be because men who chose to have a PSA test also led a healthier lifestyle and so were at lower risk of dying from prostate cancer.
The PLCO trial looked at moderate PSA testing intensity (52% in the control group) compared with higher intensity (85%, as not all men assigned to screening followed through), with no coordinated approach or treatment recommendation for abnormal results.
The ERSPC
The much bigger European trial (European Randomized Study of Screening for Prostate Cancer, or ERSPC) reported on about 162,000 men between ages 55 and 69 who were followed for an average of nine years. The men in the screening group had a PSA test every two to four years, and a result greater than 3 ng/mL was considered abnormal. In this study, there were 70% more cases of metastatic cancer and more than twice as many locally advanced cases in the control group. Balanced against this, in the PSA screening group, the chance of being diagnosed with an early cancer that would likely never lead to death was about doubled. Since a urologist interpreted and acted on any abnormal results, more men (86%) had a biopsy and many then consulted a specialist about treatment or active surveillance. We don’t yet know full details of how the control group patients were managed once they were diagnosed with cancer — it’s possible there was a bias in the management of the cancer between the two groups.
The ERSPC trial was a purer test of screening, as contamination rates are thought to be low (although not reported). It also differed from the American study in that there was follow-through on abnormal results. Even in Europe, where non-intervention is more common than in Canada, about 80% of men diagnosed with prostate cancer were treated immediately.
The ERSPC showed that men randomized to PSA screening had a 20% reduction in prostate cancer death compared to those in the control group. For those who were actually tested (because some in the screening group didn’t end up having a PSA test), the reduction was 27% (similar to that in the PLCO sub-group analysis). The study also showed that this improvement came at a high cost: 1,068 men would have to be tested and 48 treated to prevent one cancer death. Put another way, the chance of dying of prostate cancer after nine years was reduced from 0.36% to 0.29%.
What can we conclude?
Why did the ERSPC trial show a reduction in prostate cancer death and the PLCO trial not? There are several possible explanations:
- Contamination in the PLCO trial meant that there were few advanced cancers in the control (no screening) group.
- Both trials need longer follow-up (which is planned), and the results of either may change.
- The lack of coordinated follow-up and treatment for the PLCO men meant that some cancers were missed or not promptly treated.
- Men in the PLCO trial were a bit older.
- The results are actually within statistical odds of each other.
- Maybe European prostate cancer differs from that in the US (there is some evidence of American prostate cancer being less virulent)
So, does PSA screening save lives? I believe it does for men under 70 years old, but there is a big downside. Overall, PSA screening may lead to more harm than good, considering the impact on an individual of having an abnormal PSA result (= anxiety), a biopsy (unpleasant at best), followed by treatment (associated with side effects sometimes including impotence, difficulty getting life insurance, time off work, etc.) or active surveillance (deferral of treatment for about five years, at a cost of ongoing uncertainty). Of course, these risks would be well worthwhile if you were the one man in 48 to live longer as a result of therapy, but treating all the other 47 men with no survival benefit is a very high price indeed. The ERSPC study also included quality of life measures. These will be published later and will be of considerable interest, as there may be treatment benefits apart from a small chance of living longer.
The conclusion from these trials — at least until updates come out in a few years — is that men should continue to be informed about the benefits and risks of PSA testing. But it would be wrong to promote testing for everyone. Rather, the decision should be made on an individual basis following an assessment of the pros and cons. Supporters of routine PSA testing may ask the question, “Shouldn’t men have the right to know if they are at risk, and decide for themselves if they want to be treated or not?” In fact, all men are at risk of prostate cancer; unfortunately, the PSA test is not discriminating enough to sort out the important cancers from the minimal-risk cancers that would have been better left undetected. Ideally, we need a better test than PSA to distinguish those cancers that will cause death from the rest. For now, we don’t have that data, although many researchers are working on it.
Dr. Tom Pickles is a Radiation Oncologist at the British Columbia Cancer Agency and Clinical Professor at the University of British Columbia.
