Population-Based Screening for Prostate Cancer Weighing
up the Beneficial & Adverse Effects of Screening
Acta Universitatis Tamperensis No. 1782
By Tuomas Kilpeläinen
December 2012
Tampere University Press
Distributed By Coronet Books
ISBN: 9789514489730
144 pages
$82.50 Paper Original
Prostate cancer (PC) has become a major challenge for public health since a steep rise in incidence was observed in the 1990s. Screening for PC with prostate-specific antigen (PSA) has been a controversial issue for two decades, and reliable evidence for or against screening can be obtained only from large randomized trials. The European Randomized Study of Screening for Prostate Cancer (ERSPC) was commenced in the early 1990s as an international multicenter trial to assess the benefits and harms of screening for PC with PSA. The Finnish prostate cancer screening trial is the largest component of the ERSPC trial. The aim of this dissertation was to assess the prevalence of false-positive (FP) screening results both in the Finnish trial and in four other ERSPC centers, and to determine how screening affected the incidence of PC and mortality in the Finnish trial.
The Finnish population-based trial included 80,144 men. The men were randomized either to the screening arm (SA) or the control arm (CA). The men in the CA received the usual care and were not contacted. The men in the SA were tested with PSA at four-year intervals up to three times. Men with PSA ?4.0 ng/ml were referred to diagnostic work-up (prostate biopsies), as were those with PSA 3.0-3.99 ng/ml and positive ancillary test (digital rectal examination in 1996-1998; free/total PSA ratio <16% from 1999 onwards).
In Finland, 75% of the men in the SA participated at least once in screening, and the participation proportion was roughly 70% per round. In multi-center analyses, on average every tenth man attending screening was subject to an FP screening result. A quarter of men with an FP result chose not to participate in the next screening round, and of the men who did participate, approximately 50% had another FP result. The risk for a next-round PC was roughly four-fold compared to men with a previous negative screen.
In the Finnish trial, the overall incidence of PC was 8.5/1,000 person-years in the SA and 6.3/1,000 person-years in the CA (incidence rate ratio (IRR) = 1.30; 95% confidence intervals 1.24-1.36). The IRR for localized PC was 1.48 (1.40-1.56) and for advanced PC 0.74 (0.66-0.83). The cumulative mortality was 59/10,000 men in the SA and 65/10,000 men in the CA [hazard ratio (HR) 0.89 (0.75-1.07)] during 12 years of follow-up. After adjusting for bias caused by non-participation, the ratio was only slightly improved. According to a secondary analysis, a lower PSA threshold of 3.0 ng/ml would not have improved the HR substantially, whereas excluding all interval cancers would have yielded a more prominent decrease in mortality in the SA.
Screening for PC with PSA resulted in a substantial (30%) decrease in the incidence of advanced PC, but this did not transform into a statistically significant decrease in PC mortality during 12 years of follow-up. The men screened were often subject to an FP screening result and overdiagnosed cancers. The results of this dissertation do not support launching generalized population-based PC screening. It is possible that with longer follow-up the mortality reduction becomes larger, but even then cost-effectiveness and quality of life factors need to be weighed against the observed benefits.
Return to Coronet Books main page