Study design

The United States (U.S.) PLCO Trial randomized men to receive either annual screening (PSA testing for 6 years and digital rectal exam [DRE] for 4 years) or usual care, as offered by individual providers, to determine the effectiveness of screening with PSA and DRE on PCa mortality [ ]. A total of 76,693 men ages 55 to 74 at 10 U.S. study centers between 1993 and 2001 were recruited. A PSA ≥4.0 ng/mL was considered to be positive prompting a recommendation for diagnostic evaluation to both the participants and their clinicians.

Findings

At 7 years of follow-up, the incidence of prostate cancer was higher in the screening group compared to the control group (RR 1.22, [95% CI: 1.16–1.29]). Overall, similarly low rates of PCa-specific deaths occurred in both the screening and control groups at 2.0 vs. 1.7 per 10,000 person-years (RR 1.13, [95% CI: 0.75–1.70]), respectively.

At almost 17 years of follow-up, there was no statistically significant difference in PCa mortality between the 2 groups with PCa mortality rates of 5.5 and 5.9 per 10,000 person-years (RR 0.93, [95% CI: 0.81–1.08]) in the screening vs. control groups, respectively [ ]. However, a significant increase in detection of Gleason 2 to 6 disease (RR 1.17, [95% CI: 1.11–1.23]) and concomitant decrease in Gleason 8- to 10 disease (RR 0.89, [95% CI: 0.80–0.99]) was identified in the screening arm compared to usual care [ ]. Of the men who underwent at least 1 PSA screening test, 10.4% had at least 1 false positive result. Overall, 12.6% of all men in the study underwent 1 or more prostate biopsies with an overall complication rate of 2.0%, and 67.7% of all biopsies were negative [ ]. A recent secondary analysis of this initial dataset revealed that a PSA level between 0.5 to 0.99 ng/mL and between 1.0 and 1.9 ng/mL in men aged 55 to 60 were associated with 1.5% and 5.4% incidence of clinically significant PCa at 13 years, respectively [ ].

Strengths and limitations

The PLCO trial has been described as an evaluation of organized versus opportunistic screening [ ] because a high degree of contamination—i.e., PSA-testing in the control group—was noted, with very high percentages of men estimated to have undergone any PSA testing throughout the trial period (86% and 99% in the control and intervention arms, respectively) [ , ]. This diluted the power of the trial to detect a statistically significant difference in PCa mortality. In addition, rates of PSA and DRE screening in the control arm increased from 40% to 52% and from 41% to 46% between years 1 and 6, respectively. Furthermore, recommendations for prostate biopsy were not mandated by the study protocol but were at the discretion of the physician and patient. The compliance rate with biopsy was only 40.2% in the first 7-year period and 30.1% in the subsequent study period after a positive screen, which further skews interpretation of the PCa incidence rates reported. The relatively lower percentage of patients who ultimately underwent a biopsy may have also been compounded by a study-defined PSA threshold of ≥4.0 ng/mL, which is higher than that utilized in other screening trials. However, accounting for contamination and increased PCa incidence due to screening and diagnostic work-up, analytic and microsimulation models later confirmed a 27% to 32% relative risk reduction in PCa mortality in PLCO [ ].

Finally, the PLCO study revealed more Black men were diagnosed with aggressive/metastatic PCa compared to White men. However, no association was found between race/ethnicity and PCa mortality. It should be noted that the trial enrolled a very limited percentage of African American men (4.5%), so the trial was likely underpowered for subgroup analyses, representing a limitation of this study’s relatively homogenous patient population [ ].

Study design

The ERSPC trial randomized men across 7 European countries to a screening intervention arm compared to no screening to determine the effect of PSA screening on prostate cancer mortality. Those in the screening intervention group were offered PSA testing once every 4 years (every 2 years in Sweden) compared to no screening in the control group, with a PSA threshold prompting biopsy ∼3.0 ng/mL. A total of 182,000 men aged 50 to 74 were initially recruited, with 162,243 men aged 55 to 69 (core age group) included in the final analysis [ ].

Findings

At a median of 9 years of follow-up, PSA-based screening in the intervention arm reduced PCa mortality by 20% (RR 0.81, [95% CI: 0.65–0.98], P = 0.04). At the same time, the intervention arm had a higher rate of overdiagnosis of low-grade disease (72.2% vs. 54.8% Gleason 6 patients in the screening and control groups, respectively), with PCa incidence rates of 8.2% vs. 4.8% in the screening and control groups, respectively. In an intention-to-screen (ITS) analysis, the number needed to screen (NNS) was 1,410 men and 48 additional PCa cases needed to be diagnosed (NND) to prevent 1 man from dying from PCa. When analyzed per protocol, NNS decreased to 1068.

At 13 years of follow-up, PCa mortality remained divergent between the groups at 4.3 vs. 5.4 per 10,000 person-years (RR 0.79, [95% CI: 0.69–0.91)] in the screening versus control groups, respectively. NNS decreased to 781 men screened and 27 men diagnosed to prevent 1 PCa-death, with those numbers further decreasing to 570 and 18 at 16 years, respectively, with a larger absolute benefit in PCa mortality reduction with decreased overdiagnosis [ , ]. No reduction in PCa mortality was observed among men aged 70 to 74 at baseline. The incidence of metastatic disease (mPCa) was significantly lower among men assigned to the screening arm at 12 years of follow-up (RR 0.70, [95% CI: 0.60–0.82]) with NNS 323 and NND 12 to prevent 1 case of metastasis. Overall, 17.8% of men screened at least once received at least 1 false-positive PSA result, with 27.7% of screened men undergoing prostate biopsy, among whom 75.8% did not result in a PCa diagnosis [ ].

In a subsequent follow-up study by Heijnsdijk et al. analyzing the ERSPC 13-year follow-up data in a microsimulation model, a total of 73 life-years gained was reported if annual PSA screening was performed for all men 55 to 69 years of age. However, they also reported a reduction to 56 quality-adjusted life-years (QALYs) due to postdiagnosis consequences (primarily long-term side-effects from curative treatment) and 43% of cancers overdiagnosed (primarily low risk disease) [ ].

When evaluating outcomes based on ERSPC countries separately, the Finnish cohort reported a lower incidence of advanced PCa in the screening arm opposed to the control (HR 0.73, [95% CI: 0.64–0.82]), albeit no statistically significant reduction in PCa mortality (HR 0.85, [95% CI: 0.69–1.04]) [ ]. Similar findings were echoed in Spain at 15 years of follow-up; however, analysis was limited by a relatively smaller sample size and ( n = 4,276) and low overall PCa related mortality rates (0.23%) [ ]. After adjusting for contamination with 19.4% of the control group undergoing screening throughout the study period in the Rotterdam section, the benefit in mortality reduction further increased to 51% [ ].

Strengths and limitations

The ERSPC trial was a well-powered study with extended follow-up reporting a reduction in PCa-mortality, as well as in development of metastases, with implementation of routine PSA screening. Differential treatment between trial cohorts has been posited to introduce bias towards the screening arm raising concerns with internal validity, yet a more recent study reports a minimal difference of 4.2% between the arms, concluding this impact to be negligible [ , ]. However, the trial is limited by heterogeneity across the study centers in recruitment methods, screening protocols, and PSA thresholds prompting biopsy. Although the most common trigger PSA was ≥3.0 ng/mL, wide intra- and inter-site variation throughout the study period was observed.

Study design

Goteborg-1, the Swedish arm of the ERSPC trial, recruited 20,000 men aged 50-64 at baseline who were randomized to evaluate the impact of PSA screening every 2 years in the intervention group versus no screening in the control group, with trigger for biopsy prompted at a PSA value between 2.5 and 3 ng/mL. Men in the screening group were invited for biennial PSA screening up to an upper age limit of ∼70 years (range 67–71).

Findings

This study reported a favorable impact of PSA screening, with screening reducing PCa- mortality by almost half over a 14-year period. However, overdiagnosis rates remained high, with an NNS of 293 men and an NND of 12 [ ]. At 22 years of follow-up, the rate ratio for PCa mortality was 0.56 (95% CI: 0.39–0.82), favoring the screening group. A 41% reduction in PCa-mortality was reported with NNS and NND decreasing to 221 and 9, respectively [ ]. Prognostic factors associated with increased risk of PCa death were nonattendance to PSA-screening, starting screening after age 60, and stopping screening at age 70. Opportunistic screening in the control group over time did not have any effect on PCa mortality and led to more overdiagnosis [ ].

Strengths and limitations

The population-based study design initially randomized participants with subsequent invitation for participation, lending for a more representative study with appropriate generalizability.

Study design

The CAP Trial was designed to evaluate the impact of a single PSA screening test on PCa mortality [ ]. Overall, 419,582 men ages 50 and 69 from 573 primary care practices across the United Kingdom (UK) were enrolled between 2001 and 2009. The intervention group was invited to undergo a single PSA test at some point during the study period, whereas those in the control group received standard of care. A positive PSA level triggering a recommendation for a biopsy was set at ≥3 ng/mL.

Findings

Among the intervention arm, 40% ultimately attended the PSA testing clinic and 36% received a PSA test. Among those who received a PSA in the intervention arm, 11% had a PSA between 3 and 19.9 ng/mL, and 85% of this cohort ultimately underwent a prostate biopsy. Incidence of PCa was higher in the intervention vs. control groups (4.3% vs. 3.6%, RR 1.19 [95% CI: 1.14–1.25], respectively) [ ]. At a median follow-up of 10 years, no significant difference in PCa mortality was identified (RR 0.96, [95% CI: 0.85–1.08]).

Strengths and limitations

A major limitation of this study was the low rate of adherence to invitation to the intervention arm along with potential contamination due to opportunistic screening in the control arms, further diluting the impact of screening on PCa mortality [ ]. Unlike PLCO and ERSPC (including Goteborg-1), which relied on a dynamic, regular screening protocol (every 1–4 years), the intervention in CAP was a single PSA test. Parallel study designs have been attempted, including a screening study conducted in Stockholm, commencing in 1988 [ ]. Such single PSA-screening may reveal many cases of prostate cancer diagnosed at advanced stage (prevalence screen). For PSA screening to reduce PCa mortality, effective protocols have been based on repeat testing at regular intervals.