Routine blood test for prostate cancer 'doesn't save lives'

"Screening for prostate cancer does not save lives, and may do more harm than good," reports the Daily Telegraph.

"Screening for prostate cancer does not save lives, and may do more harm than good," reports the Daily Telegraph.

"Screening for prostate cancer does not save lives, and may do more harm than good," reports the Daily Telegraph.

Research involving more than 400,000 men in the UK found those invited for screening were more likely to have prostate cancer diagnosed but no less likely from die of it.

The study involved 573 GP practices, with some assigned to offer all men aged 50 to 69 a prostate specific antigen (PSA) test, while the others only offered tests if men asked for them. Men whose results suggested possible prostate cancer then had a biopsy, and those found to have cancer were treated.

The test measures the amount of PSA in the blood. Levels usually are usually higher when men have prostate cancer, but other things, such as urine infections, also raise PSA. The levels also don't tell you whether a cancer is so slow-growing it will never cause problems or if it's fast-growing and needs treatment. Fast-growing cancers can also be missed.

Furthermore, the prostate gland may get bigger as men get older, even if there are no cancerous cells in the gland. This benign prostate enlargement can also raise PSA levels.

Prostate cancer treatment can cause erection problems and urinary incontinence, so treatment is usually only recommended for more aggressive forms.

But this doesn't mean older men should ignore symptoms potentially associated with prostate cancer – these usually involve problems with urination, such as a frequent or sudden urge to pee. Contact your GP for advice if you experience any symptoms like this.

Where did the story come from?

The research team included members from the University of Bristol; University Hospitals Bristol NHS Trust; Hull York Medical School; Royal United Hospitals Bath; the University of Oxford; Bristol, North Somerset and South Gloucestershire Clinical Commissioning Groups; and the University of Cambridge.

The study was funded by Cancer Research UK and the National Institute of Health Research, and published in the peer-reviewed Journal of the American Medical Association.

It was covered widely in the UK media. Most reports were balanced and reasonably accurate, rightly making the point that a PSA test carried out in isolation is unlikely to be of much practical use.

What kind of research was this?

This was a cluster randomised controlled trial, with randomisation carried out at general practice rather than individual participant level. Randomisation of this type is usually a good way to see what effect a test or treatment has.

What did the research involve?

Researchers selected GP surgeries within geographic areas near participating hospitals, and randomly assigned them to either the screening or control group. They then approached the practices to see if they wanted to take part. More practices agreed to be in the control group (302) than the screening group (271).

To be eligible for the study, men had to be between 50 and 69, and not have been previously diagnosed with prostate cancer. There were 189,386 men in the screening group and 219,439 in the control group.

Eligible men who were enrolled at practices in the screening group were sent an invitation to have a single PSA test. Those with a PSA level of over 3ng/mL – considered a raised level in men aged 50 to 69 – were offered a biopsy and then prostate cancer treatment if the biopsy showed cancerous cells.

Men in the control group were not offered screening but were able to request a PSA test if they wanted one, as is standard practice in the UK.

All the men in the study were followed up for an average of 10 years to see if they had been diagnosed with prostate cancer and if they had died of prostate cancer.

The researchers compared diagnosis and death rates between the men who were offered screening and those who were not. They also looked at the stage of cancers diagnosed in the groups.

What were the basic results?

Men in the screening group were more likely to be diagnosed with cancer within 10 years of the test. Specifically:

  • 67,313 men in the screening group (36%) attended the clinic and had a PSA test
  • 11% of those tested had a raised PSA level, of which 85% had a biopsy
  • 8,054 men in the screening group (4.3%) were diagnosed with prostate cancer, compared with 7,853 (3.6%) of those in the control group

However, there was no difference between the screening group and the control group in the rates of prostate cancer deaths after 10 years – about 3 in every 1,000 died of prostate cancer in both groups. This implies the PSA test screening did not achieve its aim of diagnosing fast-growing cancers in time to treat them and prevent fatality.

The results suggest 3 main reasons for this.

Firstly, more early-stage cancers, which were less dangerous and possibly less likely to grow, were diagnosed among men in the screening group than the control group.

Furthermore, of the 549 men in the screening group who died of prostate cancer, 68 (12.4%) had low PSA levels at screening so did not have a follow-up biopsy or treatment.

Finally, some men were seriously harmed by treatment. There were 8 deaths in the screening group related to either the biopsy or prostate cancer treatment and 7 in the control group. The study did not record other potential harm from treatment, such as the well-known problems with incontinence and sexual function.

How did the researchers interpret the results?

The researchers said that longer-term follow-up of their figures was ongoing but that the findings "do not support single PSA testing for population-based screening".

In a press release issued by Cancer Research UK, one of the researchers said they now must find "better ways" to diagnose fast-growing cancers that need early treatment.


This research is valuable in the debate over whether routine prostate cancer screening using the PSA test should be made widely available. Based on this study, the answer is clearly no: using the test to screen for prostate cancer in this way does not help – and it may even harm.

New research is looking at ways to make the PSA test more accurate, but it may still miss some fast-growing cancers, as it did in this study. Researchers are also looking at using MRI to improve the accuracy of biopsies, but these scans are only done after a high-PSA test result.

The research did have some limitations.

Although the study was large, only 36% of those in the screening group actually had a PSA test. It's possible the men who attended screening may have been more concerned about their health in general so were also more likely to have a healthy lifestyle. However, that would usually mean they would be less likely to die of prostate cancer, but the results do not bear that out.

The study reported results after 10 years. Because prostate cancer grows slowly in most cases, this might be too soon to see the full effect of early screening. The researchers are continuing to follow the men up, so it will be interesting to see the results after 15 years.

Men were offered only one PSA test, whereas some previous studies have offered repeated tests every few years. It's possible repeated screening might have picked up some of the fatal cancers missed after one test. However, this must be balanced against the overdiagnosis of slow-growing cancers from repeated rounds of screening.

If you're worried about your prostate cancer risk – for example, because you have a family history of it – talk to your GP about your individual risk. If you are over 50 and decide after discussion with your GP to have a PSA test, you can have one free on the NHS.

Find out more about the PSA test.

Article Metadata Date Published: Wed, 7 Mar 2018
Author: Zana Technologies GmbH
NHS Choices