New studies examine effect of PSA screening on prostate cancer death rates

In collaboration with the Press Association

Two studies published in the latest issue of the New England Journal of Medicine have examined the impact of prostate-specific antigen (PSA) screening on death rates from prostate cancer.

The PSA test measures the levels of PSA protein in the blood. This protein can indicate the presence of prostate cancer, although some men with the disease have normal PSA levels while many men with high PSA do not have cancer.

The PSA test has been used for about 20 years and has been shown to help diagnose prostate cancer at an earlier stage. The test is widely used in the US as part of men's health insurance check-ups. British men are able to request a PSA test from their GP as part of the Prostate Cancer Risk Management Programme, although there is no national prostate cancer screening programme in place.

 

However, it is not yet clear whether the use of the PSA test reduces the actual number of deaths from prostate cancer. There are also concerns that the test may cause many false alarms and leads to excessive treatment of men whose prostate cancers would not have affected them during the natural course of their lifetime if left untreated.

Interim results from two new studies have now been published in the New England Journal of Medicine which looked at the impact of PSA testing on large numbers of men.

The first - the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (PLCO) - is being conducted in the US and enrolled around 76,000 men between 1993 and 2001.

The study, which has followed participants for an average of 11 years, has so far found evidence of fewer deaths amongst men offered screening, although some experts have pointed out limitations in the study design.

However, the second study, which is being conducted in Europe and involves a much larger number of men between the ages of 55 and 69, provides some support for the routine use of PSA testing.

Around 162,000 men were recruited for the European Randomised Study of Screening for Prostate Cancer (ERSPC) trial and assigned to screening with PSA every four years or regular digital rectal exams. They have so far been followed for an average of nine years.

Around twice as many cases of prostate cancer have been diagnosed in the PSA screening group as in the non-screening group, interim results show.

PSA screening also appears to be associated with a 20 per cent reduction in mortality, with 326 patients in the non-screening group dying from prostate cancer compared with just 214 men in the screening group.

The European study suggests that one prostate cancer death could be prevented for every 1,410 men screened and 48 men treated during a ten-year period.

However, the results also show that many men may have undergone unnecessary biopsies as a result of PSA testing. Among the 73,000 men who were screened, more than 17,000 biopsies were performed - many more than in the non-screening group.

Writing in an accompanying editorial in the New England Journal of Medicine, Dr Michael Barry, from Massachusetts General Hospital and Harvard Medical School, concludes that routine PSA screening has a "modest effect" at best on prostate cancer mortality during the first ten years of follow-up, and that this benefit "comes at the cost of substantial over-diagnosis and overtreatment".

"It is important to remember that the key question is not whether PSA screening is effective, but whether it does more good than harm," he noted.

"Some well-informed clinicians and patients will still see these trade-offs as favourable, others will see them as unfavourable. As a result, a shared decision-making approach to PSA screening, as recommended by most guidelines, seems more appropriate than ever."

Dr Phil Kantoff, professor of medicine at Harvard Medical School, took part in a discussion of the papers which was hosted by the journal.

He noted that the number of deaths during the first ten years after screening is "quite modest" and that men who are not expected to survive for longer than ten years regardless of whether or not they have prostate cancer "could conceivably forgo screening".

He continued: "I do think that there is going to turn out to be a reduction in mortality associated with PSA-based screening, but what I also firmly believe is that not everybody diagnosed with prostate cancer needs to be pigeonholed into a treatment paradigm.

"We need to individualise, because clearly there are many patients who are diagnosed with prostate cancer that do not need to be treated, can be observed safely, and will not die of their cancer."

Commenting on the European study, Professor Peter Johnson, Cancer Research UK's chief clinician, noted: "It's very difficult to prove, even in large trials like this, that PSA screening has a direct impact on reducing cancer deaths, so we need to see what the results will be with longer follow up.

"For some men, detecting prostate cancer early through screening can be lifesaving, while for others it will lead to unnecessary treatment and side-effects like impotence and incontinence. But the important thing is that men in their 50s and 60s discuss this with their doctor and feel able to request a PSA test easily."

Harpal Kumar, Cancer Research UK's chief executive, added: "Despite the risks of unnecessary treatment, more men as a result of this study will want a PSA test. The government should begin a feasibility study to assess the implications for screening and treatment.

"We still don't know what the best treatment approach is for early disease, so it's important we find answers to this as soon as possible through research currently funded by Cancer Research UK and others. Scientists also need to accelerate their efforts to find markers to distinguish between slow growing and aggressive forms of prostate cancer so that we know which ones to treat and which ones are best to monitor."

References

Schroder, F.et al (2009). Screening and Prostate-Cancer Mortality in a Randomized European Study New England Journal of Medicine DOI: 10.1056/NEJMoa0810084