PSA Screening: Why ERSPC and PLCO Reached Different Conclusions

The European Randomized Study of Screening for Prostate Cancer, or ERSPC, found that invitation to repeated PSA screening reduced prostate cancer mortality by about one fifth in its core age group. The US Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial, or PLCO, found no significant mortality difference. The apparent conflict is largely a difference in what the control groups received.

In ERSPC, the contrast was closer to organized screening versus little screening. In PLCO, PSA testing was already common before enrollment and became very common in usual care: an estimated 78% of men in the PLCO control group had at least one PSA test during the screening phase. PLCO therefore tested an organized, more intensive screening program against opportunistic screening, which leaves less separation for a mortality benefit to appear.

Reconstruct the intervention, not just the trial name#

A screening trial is a comparison of pathways. Invitation is followed by attendance, testing, repeat testing, diagnostic workup, biopsy, cancer classification, surveillance or treatment, and long-term follow-up. A difference at any point can change mortality.

ERSPC enrolled men across several European countries, with a prespecified core group aged 55 to 69. Screening intervals and biopsy thresholds varied by center, although PSA testing every two to four years was typical, and men in the control group were not invited into the organized program.

PLCO randomized more than 76,000 US men aged 55 to 74. The intervention offered annual PSA testing for six years and digital rectal examination for four years; further diagnostic workup was managed in usual care rather than dictated by the trial. Many participants had already been tested before enrollment, and PSA use continued outside the intervention arm.

Randomization remained valid in both trials. The important issue is estimand: ERSPC estimated the effect of an organized invitation against relatively limited background testing, while PLCO estimated the added effect of an intensive program in a setting where opportunistic testing was already widespread.

What ERSPC found#

The first ERSPC mortality report in 2009 found a relative reduction in prostate cancer death of about 20% among men in the core age group, alongside substantially more prostate cancer diagnoses in the screening group. Longer follow-up strengthened the evidence that the mortality effect was real, while showing that the absolute difference remained modest.

The USPSTF's 2018 synthesis estimated that, over about 13 years, PSA-based screening might prevent approximately 1.3 prostate cancer deaths per 1,000 men screened and about 3 cases of metastatic prostate cancer per 1,000. Those averages are easier to interpret than the relative reduction alone. A 20% relative reduction can sound large even when the starting risk is low.

Results also varied across ERSPC centers. Screening intervals, biopsy thresholds, compliance, and subsequent treatment differed. Prostate cancer mortality was significantly reduced in some centers but not others. A pooled result is informative, but it does not turn every implementation into the same intervention.

What PLCO found#

After nearly 15 years, PLCO found no significant difference in prostate cancer mortality. The USPSTF report summarized rates of 4.8 versus 4.6 deaths per 10,000 person-years, with a relative risk of 1.04 and a 95% confidence interval from 0.87 to 1.24.

That result shows no added mortality benefit from the PLCO screening program over the screening that occurred in its usual-care arm, and it does not provide a clean estimate of PSA screening versus never screening. Men in the intervention group averaged about five PSA tests during the screening phase, while controls averaged about three. When both groups receive much of the active ingredient, the intention-to-treat contrast shrinks.

Control contamination is not a reason to discard the trial. It accurately reflects a policy question relevant to a high-testing health system: does a formal annual program add benefit beyond common opportunistic testing? Its answer can coexist with ERSPC's answer to a different contrast.

Other design differences also matter#

Control-group testing is the central reconciliation, but not the only one. The trials differed in age ranges, screening intervals, PSA thresholds, prior testing, attendance, biopsy practice, and treatment after diagnosis. Lead time and length bias shape both studies.

Lead time makes survival from diagnosis appear longer simply because screening starts the clock earlier. Disease-specific mortality avoids that particular illusion, which is why it is the central endpoint; length bias means repeated screening is more likely to detect slower-growing tumors that remain detectable for longer. Those tumors may never have caused symptoms.

Treatment can modify the outcome after detection. If treatment choice or quality differs between randomized groups, part of a mortality difference may reflect downstream care. The screening test cannot be evaluated as though biopsy and treatment were identical, automatic steps.

A modeling analysis published in 2017 harmonized the intensity of screening received in ERSPC and PLCO, and it concluded that both trials were compatible with a reduction in prostate cancer mortality when the different screening contrasts were accounted for. This is a reasoned reconciliation based on assumptions, not a new randomized comparison.

The harm pathway starts before treatment#

PSA is prostate-specific, not cancer-specific. Benign prostatic enlargement, inflammation, infection, recent instrumentation, and ordinary biological variation can raise it. In ERSPC sites reporting false positives, roughly one in six screened men had at least one false-positive result, and about two thirds of positive results in the first round were false positives.

A positive PSA can lead to repeat blood tests, imaging, biopsy, anxiety, and complications. Biopsy can cause pain, bleeding, infection, and occasional hospitalization. Newer MRI-based diagnostic pathways may reduce some unnecessary biopsies, but the landmark mortality trials largely predated their current use.

Overdiagnosis is different from a false positive. It means finding a genuine prostate cancer that would not have caused symptoms or shortened life. Because screening preferentially detects slow-growing disease, some diagnosed men face years of surveillance or treatment for a cancer that would never have become clinically important.

Treatment can cause urinary incontinence, erectile dysfunction, bowel symptoms, and other long-term effects. Active surveillance reduces immediate overtreatment, but it adds repeat testing and does not eliminate later treatment. The decision you are actually making covers the whole downstream pathway, not the blood test in isolation.

Current USPSTF interpretation#

As of July 15, 2026, the USPSTF lists its prostate cancer screening recommendation as being updated, but the 2018 final recommendation remains the current final statement. It assigns grade C to PSA-based screening for men aged 55 to 69. Grade C means the net benefit is small and the choice should reflect your own circumstances and preferences. It assigns grade D to screening in men aged 70 or older, concluding that harms outweigh benefits on average in that age group.

The Task Force recognizes higher prostate cancer incidence and mortality among Black men and higher risk with a family history, particularly relatives with aggressive or fatal disease. Yet randomized evidence is limited for defining different starting ages or intervals in these groups. Higher baseline risk can increase absolute potential benefit, while evidence gaps make precision difficult.

These recommendations are time-stamped. Other professional organizations use different age windows and risk frameworks, and an in-progress USPSTF review may eventually change the final statement. Any article quoting a screening grade should show its date and issuing organization.

How to read a screening statistic#

Ask what denominator and time frame it uses. “One death prevented per 1,000 screened over about 13 years” is not the same as one per 1,000 cancers found. Relative mortality reduction, absolute mortality reduction, number invited, number actually screened, number biopsied, and number treated answer different questions.

Also separate disease-specific from all-cause mortality. Neither ERSPC nor PLCO established an all-cause mortality benefit. Prostate cancer deaths are a small portion of all deaths in the enrolled age groups, so even a real disease-specific effect may not visibly shift all-cause mortality.

A shared-decision evidence frame#

A good discussion with your clinician covers:

• baseline risk based on age, family history, ancestry, and prior PSA pattern;
• the small average absolute chance of preventing metastasis or prostate cancer death;
• the probability of false-positive testing and diagnostic procedures;
• the possibility of overdiagnosis and prolonged surveillance;
• treatment effects that matter to urinary, sexual, and bowel function;
• other illnesses and life expectancy, which alter the time available to benefit;
• your own preference for early information versus avoiding downstream intervention.

This is why opposite personal choices can both be evidence informed. The trials quantify population averages; they do not assign a value to the tradeoff for you.