Loading...
Researchers and clinicians have long awaited results from two randomized trials that were supposed to tell us, once and for all, whether prostate-specific antigen (PSA) screening is beneficial. In 2009, the first round of results was published, leaving us with more questions than answers (JW Gen Med Mar 18 2009).
In the PLCO trial, 77,000 American men (age range, 55–74) were randomized to annual screening (PSA testing plus annual digital rectal examinations) or to no screening for 6 years. During 10 years of follow-up, researchers found no difference in prostate cancer–related deaths (roughly 85 in each group). However, during each of the 6 screening years, 40% to 50% of the control group received some form of PSA testing outside of the trial.
In the ERSPC trial, 182,000 men (age range, 50–74) in seven European countries were randomized to PSA screening (with various intervals and follow-up protocols) or to no screening. During a mean follow-up of 9 years, fewer prostate cancer–related deaths occurred in the screened group than in the control group — a difference of 7 per 10,000 screened men (P=0.04).
Publication of these studies was followed by several letters to the editor of the New England Journal of Medicine from readers who charged that both studies had various theoretical and methodological problems, including inadequate follow-up, “contamination” among controls (PSA testing outside the trials), and misclassification of deaths. However, study authors did provide appropriate responses to these concerns (N Engl J Med 2009; 361:202).
Results from several observational studies added fuel to the fire. One study demonstrated substantial fluctuation of PSA levels from year to year, with levels ≥2.5 ng/mL often dipping below this threshold on subsequent annual testing (JW Gen Med May 19 2009). And, in a large epidemiological study, researchers compared the incidence of prostate cancer before 1987 with that in the PSA era (1987–2005). Annual incidence rose dramatically during the early years of PSA screening and stayed high in later years — contrary to what would be expected if PSA testing identified only clinically significant cancers. In a worst-case scenario, the authors estimated that as many as 1 million men were treated unnecessarily during these 2 decades (JW Gen Med Sep 24 2009).
In a related study, 10-year prostate cancer–specific mortality in older men was relatively low for highly and moderately differentiated cancers (8% and 9%) found during the early era of PSA screening (1992–2002) and was roughly 60% lower than in historical controls. One explanation for this finding is that PSA screening identifies less-aggressive cancers for which outcomes are better regardless of treatment (JW Gen Med Oct 15 2009).
Unfortunately, the conflicting initial findings from the PLCO and ERSPC trials have not settled the PSA controversy. PSA testing is neither highly sensitive nor specific, and even urologists who advocate screening acknowledge that overdiagnosis and overtreatment are major problems (J Urol 2009; 182:2232). Perhaps longer-term follow-up from these two trials will settle the issue — but I wouldn't count on it.