Prostate Cancer Screening: PSA Testing, Benefits, Risks and What the Evidence Shows
In brief
PSA screening can prevent some deaths from prostate cancer.
In a large European trial that began in the 1990s, inviting 1,000 men aged 55–69 to a repeated PSA-screening programme prevented about 2 prostate-cancer deaths over 23 years.
Also, in an analysis from the same European trial, over 12 years there were about 3 fewer cases of prostate cancer that had spread to other parts of the body per 1,000 men invited to repeated PSA screening, compared with the trial's control group. These are not an additional 3 men on top of the 2 deaths prevented. They are a different measure of screening's benefit, measured over a different period.
But testing can start a chain of further investigation. A raised PSA can lead to MRI, biopsy and a prostate-cancer diagnosis. Some cancers found through screening would never have caused symptoms or shortened the man's life. Finding them can lead to years of surveillance and, in some cases, treatment that was not needed.
The screening pathway has changed substantially since the major trials began. MRI can now help determine who needs a biopsy, and men with lower-risk cancers can often be monitored rather than treated immediately. This appears to reduce some of the harms associated with older screening pathways.
But there is an important gap in the evidence.
We have long-term evidence that older PSA screening programmes can prevent some prostate-cancer deaths. We do not yet have equally long-term evidence showing the complete balance of benefits and harms when today's MRI-based pathway is used.
What does PSA screening actually mean?
PSA stands for prostate-specific antigen, a protein produced by the prostate. A PSA test measures its level in the blood.
A raised PSA does not mean that a man has prostate cancer. PSA can rise for several reasons, including enlargement or inflammation of the prostate. But prostate cancer can sometimes be present when PSA is not particularly high.
Screening means testing someone who does not have symptoms, with the aim of finding prostate cancer earlier, before it has grown or spread.
This is different from using PSA to investigate symptoms.
It is also important to understand that not all prostate cancers behave in the same way.
Doctors and researchers often use the term “clinically significant prostate cancer” for cancers considered more likely to grow, spread or cause harm. Other, lower-risk prostate cancers may grow so slowly that they would never cause symptoms or shorten a man's life.
That distinction lies at the heart of the PSA-screening debate: screening is intended to find clinically significant cancers earlier, but it can also find cancers that would never have harmed the man.
Can PSA screening prevent deaths from prostate cancer?
Yes. But it is important to look at the absolute numbers.
The largest randomised evidence comes from the European Randomized Study of Screening for Prostate Cancer (ERSPC).
The trial began in the 1990s and involved more than 160,000 men, with the main analysis concentrating on men aged 55–69. Men in the screening group were invited for repeated PSA testing. Men in the control group were not invited to the organised screening programme.
After about 23 years of follow-up, prostate-cancer mortality was approximately 13% lower among men invited to screening than in the control group.
In absolute terms, the difference was about 0.22 percentage points. That is equivalent to approximately 2.2 fewer prostate-cancer deaths for every 1,000 men invited to screening, compared with the control group.
Or, expressed another way, about 456 men needed to be invited to screening to prevent one prostate-cancer death over that period.
At the same time, prostate cancer was diagnosed about 30% more often among men invited to screening than in the control group.
Finding more cancers is not necessarily a bad thing: some of those cancers are the ones screening is intended to find early. But screening can also find cancers that would never have caused symptoms or shortened a man's life.
This is known as overdiagnosis, and we will return to it below.
Did screening also reduce the number of cancers that spread?
Yes.
An analysis involving 76,813 men aged 55–69 from four ERSPC centres looked at prostate cancer that had spread to other parts of the body — known as metastatic prostate cancer.
After about 12 years of follow-up, there were approximately 3.1 fewer cases of metastatic prostate cancer per 1,000 men invited to screening, compared with the control group.
This is a different outcome, measured over a different period, from the 23-year prostate-cancer mortality result. The two figures should therefore not be added together.
The likely explanation is straightforward: screening finds some cancers earlier, allowing treatment before they have had an opportunity to spread.
What happened in the large UK trial?
The UK CAP trial provides a different test of PSA screening.
It involved more than 415,000 men aged 50–69.
Unlike ERSPC, the intervention was not repeated screening. Men were offered a single invitation to have a PSA test.
After 15 years, prostate-cancer mortality was:
6.9 deaths per 1,000 among men invited for PSA testing
7.8 deaths per 1,000 in the control group.
That is a difference of about 0.9 prostate-cancer deaths per 1,000 men invited.
But CAP needs to be interpreted carefully.
Only about 40% of the men invited actually attended for PSA testing, while some men in the control group obtained PSA tests outside the trial.
So CAP is best understood as testing the effect of a policy of offering one PSA test, rather than comparing men who were tested with men who were never tested.
Does preventing prostate-cancer deaths mean men live longer?
This is a surprisingly difficult question.
Randomised trials show a reduction in prostate-cancer mortality. They have not demonstrated a reduction in all-cause mortality — deaths from any cause.
In CAP, for example, after 15 years about 23.2% of men in the screening group and 23.3% in the control group had died from any cause.
That does not prove that PSA screening adds no life.
If screening prevents a prostate-cancer death, the man concerned may live for additional years before eventually dying from something else. Against a background of many deaths from other causes, however, that difference can be extremely difficult for a trial to detect.
A 2012 modelling study using ERSPC data projected outcomes over men's remaining lifetimes, assuming annual PSA screening between ages 55 and 69. It estimated approximately 8.4 years of additional life for each prostate-cancer death prevented.
But that was a modelled lifetime projection, not something directly observed in the trial. Its results depended on assumptions about future screening, diagnosis, treatment and quality of life, using evidence from an earlier era.
So the most accurate conclusion is:
Randomised trials show that PSA screening can reduce deaths from prostate cancer. They have not demonstrated a reduction in overall mortality.
What can happen because you have a PSA test?
The possible harm of PSA screening is not primarily the blood test itself.
It is what may happen afterwards.
A raised PSA can lead to further PSA tests, MRI, biopsy and possibly a prostate-cancer diagnosis.
A biopsy can cause complications including bleeding, infection and urinary retention.
Biopsy techniques have changed. In a recent randomised trial using a modern biopsy technique, for example, there were no infections among 372 men, while urinary retention requiring intervention occurred in about 0.3%.
Those figures should not be treated as universal biopsy complication rates, but they show that some of the risks associated with older biopsy techniques can now be reduced.
But the larger screening problem is not simply biopsy complications.
It is finding cancers that would never have caused symptoms or shortened a man's life.
What is overdiagnosis?
Overdiagnosis means finding a prostate cancer through screening that would never have caused symptoms or shortened that man's life if it had remained undiscovered.
It does not mean the diagnosis was wrong.
The cancer is real.
The problem is that there is no way, at the moment of diagnosis, to know with certainty what would have happened to that particular cancer during the rest of that man's life.
This is especially important in prostate cancer because some cancers grow very slowly.
The exact amount of overdiagnosis caused by PSA screening is difficult to establish. It depends on a man's age, the screening strategy, how cancers are investigated and classified, and how long men are followed.
Lifetime overdiagnosis cannot simply be observed directly, because we cannot know what would have happened to an individual cancer had it never been discovered. Estimates therefore depend partly on modelling.
A modelled illustration
The UK National Screening Committee has published a modelled illustration of what might happen if 1,000 men aged 50–60 underwent a one-off PSA test.
The model estimated that this could result in approximately:
28 prostate-cancer diagnoses
up to 20 cancers being overdiagnosed
12 men receiving unnecessary treatment
up to 2 men's lives being extended.
These are modelled estimates, assembled using evidence from several studies and models. They are not the observed results of 1,000 men followed in a single trial, and they should not be treated as universal figures for every age or screening strategy.
They do, however, illustrate the central trade-off in prostate-cancer screening.
Screening may find a cancer early enough to prevent death in some men. It can also find cancers in other men who would never have been harmed by them.
Has MRI changed this?
Yes — substantially.
The major mortality trials began before today's MRI-based diagnostic pathway existed.
Historically, a raised PSA was much more likely to lead directly to systematic prostate biopsy.
Today, MRI can be used after a raised PSA to help identify which men are more likely to have clinically significant prostate cancer and which areas of the prostate should be biopsied.
This can reduce unnecessary biopsies and reduce the detection of lower-risk cancers.
The Göteborg-2 trial gives an indication of the effect.
After four years, among men aged 50–60, an MRI-targeted biopsy strategy was associated with 57% fewer diagnoses classified as clinically insignificant than a strategy that included systematic biopsy.
Here, clinically insignificant does not mean “not cancer”. It means cancer classified as sufficiently low-risk that detecting it through screening is less likely to benefit the man.
But the 57% figure does not mean MRI reduced lifetime overdiagnosis by 57%.
Nor has the trial yet established that reducing systematic biopsy in this way can preserve the long-term benefits of screening without missing clinically significant cancers that would eventually cause harm.
Longer follow-up is needed.
Does a prostate-cancer diagnosis always mean treatment?
No.
This has also changed.
Men diagnosed with lower-risk prostate cancer can often undergo active surveillance rather than immediate surgery or radiotherapy.
Active surveillance means monitoring the cancer and treating it later if there are signs that it is becoming more concerning.
This matters because treatment itself can have substantial long-term effects.
The UK ProtecT trial followed men with localised prostate cancer who were allocated to active monitoring, surgery or radiotherapy.
After 15 years, prostate-cancer mortality was low in all three groups:
3.1% with active monitoring
2.2% with surgery
2.9% with radiotherapy.
There was no statistically significant difference in prostate-cancer mortality between the groups.
But there was a difference in progression.
Prostate cancer spread to other parts of the body in:
9.4% of men assigned to active monitoring
4.7% assigned to surgery
5.0% assigned to radiotherapy.
So treatment reduced progression and spread, even though the trial had not demonstrated a mortality advantage at 15 years.
Treatment can also cause lasting harm.
Between seven and twelve years after treatment, approximately 18–24% of men in the surgery group reported needing pads for urinary leakage, compared with about 3–8% after radiotherapy and 9–11% in the active-monitoring group.
There is another important qualification.
ProtecT began many years ago, and its active-monitoring protocol was not the same as modern active surveillance. Many men initially assigned to monitoring eventually received surgery or radiotherapy.
So ProtecT helps us understand the consequences of different approaches to localised prostate cancer, but it is not a perfect description of what happens under today's pathway.
Why does age matter?
Age changes both sides of the calculation.
The strongest randomised evidence for PSA screening comes mainly from men aged approximately 55–69.
As men get older, prostate cancer becomes more common.
But so does the chance of finding a slow-growing cancer that would never have caused symptoms during the man's remaining lifetime.
That means the potential for overdiagnosis generally rises with age.
At the same time, evidence that routine screening prevents prostate-cancer deaths becomes less certain at older ages because older men were less well represented in the major screening trials and have a greater chance of dying from other causes before a screen-detected cancer would have caused harm.
This is why a PSA result cannot sensibly be interpreted without considering age and the wider clinical context.
Why doesn't the NHS screen every man?
The UK does not currently have a population PSA-screening programme for all men.
In 2026, the UK National Screening Committee concluded that population screening for prostate cancer “is more likely to cause more harm than good.”
That conclusion reflects a population-policy judgement about the balance between deaths prevented and the harms created by screening millions of men.
The committee did, however, recommend targeted PSA screening every two years between ages 45 and 61 for a much smaller group of men who have a pathogenic BRCA2 genetic variant together with a relevant family history.
For men with a family history but without the relevant BRCA2 variant, the committee concluded that screening was more likely to cause more harm than good.
For Black men, who have a higher risk of prostate cancer, the committee concluded that the available evidence was not sufficient to establish that the benefits of targeted screening outweigh the harms.
Research continues, including the large UK TRANSFORM study, which is investigating newer approaches to prostate-cancer screening.
There is an important distinction here.
Whether the NHS should systematically invite millions of men for PSA screening is a population-policy question.
Whether an individual man wants to have a PSA test is a different question.
The evidence relevant to those two decisions overlaps, but the questions are not identical.
What don't we know?
This may be the most important part of the story.
The evidence comes from two different eras.
Older screening trials have long enough follow-up to show that repeated PSA screening can prevent some prostate-cancer deaths.
Newer studies show that using MRI before biopsy can reduce unnecessary biopsies and the diagnosis of lower-risk cancers.
But we do not yet have the same long-term evidence for today's screening pathway.
So the important unanswered question is:
Can today's MRI-based approach retain the life-saving benefit of PSA screening while reducing overdiagnosis and unnecessary treatment?
We do not yet know.
That is the missing piece.
The bottom line
PSA screening can prevent some deaths from prostate cancer.
In a large European trial that began in the 1990s, inviting 1,000 men aged 55–69 to a repeated screening programme prevented about 2 prostate-cancer deaths over 23 years. Screening can also reduce the number of men whose prostate cancer spreads to other parts of the body.
But the trials have not shown that PSA screening reduces the overall risk of dying from any cause.
Screening also finds more prostate cancers. Some of these would never have caused symptoms or shortened a man's life, potentially leading to unnecessary investigation, surveillance or treatment.
Modern MRI, more selective biopsy and active surveillance can reduce some of these harms.
But today's screening pathway has not yet been followed for long enough to establish its complete long-term balance of benefits and harms.
So the evidence shows both a real but relatively small reduction in prostate-cancer deaths and a real risk of overdiagnosis and unnecessary treatment. What remains uncertain is exactly how that balance changes when PSA testing is followed by today's MRI-based pathway.