Correspondence

Prostate-Cancer Mortality after PSA Screening

N Engl J Med 2012; 366:2228-2231June 7, 2012DOI: 10.1056/NEJMc1204298

Article

To the Editor:

In their article, Schröder et al. (March 15 issue)1 provide an update on the European Randomized Study of Screening for Prostate Cancer (ERSPC) after 11 years of follow-up. Their extended analysis is of great importance given the considerable controversy regarding the benefits, risks, and costs of population-based screening for prostate cancer, as reflected in the lively discussion that followed the release of the draft recommendation against screening by the U.S. Preventive Services Task Force.2 It is an established principle of evidence-based medicine to base clinical decision making on the entire body of evidence rather than on individual trials. We recently published the results of a meta-analysis3 of prostate-cancer screening trials that included data on 302,500 patients. After updating our analysis with the results of six randomized, controlled trials involving 329,643 patients, including the most recent data from Schröder et al., we found the pooled relative risk of death from prostate cancer to be 0.98 (95% confidence interval [CI], 0.83 to 1.15), which fails to indicate that population-based screening for prostate-specific antigen (PSA) is beneficial. (For more information, see the figure in the Supplementary Appendix, available with the full text of this letter at NEJM.org.)

Mia Djulbegovic, B.S.
University of South Florida, Tampa, FL

Molly M. Neuberger, B.S.
Philipp Dahm, M.D., M.H.Sc.
University of Florida, Gainesville, FL
p.dahm@urology.ufl.edu

No potential conflict of interest relevant to this letter was reported.

4 References

1. 1

   Schroder FH, Hugosson J, Roobol MJ, et al. Prostate-cancer mortality at 11 years of follow-up. N Engl J Med 2012;366:981-990
   Free Full Text | Web of Science | Medline

2. 2

   Screening for prostate cancer: draft recommendation statement. Rockville, MD: U.S. Preventive Services Task Force (http://www.uspreventiveservicestaskforce.org/uspstf12/prostate/draftrecprostate.htm).
   

3. 3

   Djulbegovic M, Beyth RJ, Neuberger MM, et al. Screening for prostate cancer: systematic review and meta-analysis of randomised controlled trials. BMJ 2010;341:c4543-c4543
   CrossRef | Web of Science

4. 4

   Andriole GL, Crawford ED, Grubb RL III, et al. Mortality results from a randomized prostate-cancer screening trial. N Engl J Med 2009;360:1310-1319[Erratum, N Engl J Med 2009;360:1797.]
   Free Full Text | Web of Science | Medline

To the Editor:

Nonattendance, contamination, and prerandomization screening can reduce the power of a randomized screening study to ascertain whether a true reduction in disease-specific mortality exists when men undergo randomization to actual care or screening. These factors affected both the PLCO1 and the ERSPC randomized prostate-cancer screening studies, but to different degrees, and can explain the different results. Although estimated rates of nonattendance in the two studies were similar (15.0% in the PLCO trial and 17.4% in the ERSPC) (Table 1Table 1Comparison of Randomized Prostate-Cancer Screening Trials.), estimated contamination rates differed substantially, with a rate of 85.0% in the PLCO trial2 and 24.0% in the Rotterdam cohort of the ERSPC.3 In addition, only the ERSPC was powered to overcome the contamination rate. With regard to prerandomization screening for PSA, 44.0% of all participants in the PLCO trial reported that they had undergone previous screening.1 This information was not reported in the ERSPC, but a low rate would be expected because the study was conducted with a PSA-naive population. Finally, given that 85.0% of men in both randomized groups in the PLCO trial underwent PSA screening at least once, the validity of the randomization comes into question. Therefore, when making recommendations on guidelines for PSA testing, we should remember that the difference in the rate of PSA testing between those who underwent randomization to usual care and those who underwent randomization to testing in the ERSPC, which concluded that there was a benefit from testing, was 58.6%, whereas in the PLCO study, which concluded that there was not a benefit, the difference in the rate of PSA testing was 0%.

Anthony V. D'Amico, M.D., Ph.D.
Brigham and Women's Hospital, Boston, MA
adamico@lroc.harvard.edu

No potential conflict of interest relevant to this letter was reported.

3 References

1. 1

   Andriole GL, Crawford ED, Grubb RI III, et al. Prostate cancer screening in the randomized Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial: mortality results after 13 years of follow up. J Natl Cancer Inst 2012;104:125-132
   CrossRef | Web of Science

2. 2

   Pinsky PF, Black A, Kramer BS, Miller A, Prorok P, Berg C. Assessing contamination and compliance in the prostate component of the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial. Clin Trials 2010;7:303-311
   CrossRef | Web of Science

3. 3

   Kerkhof M, Roobol ML, Cuzick J, et al. Effect of the correction for noncompliance and contamination on the estimated reduction of metastatic prostate cancer within a randomized screening trial (ERSPC section Rotterdam). Int J Cancer 2010;127:2639-2644
   CrossRef | Web of Science

To the Editor:

In the ERSPC, after 11 years follow-up, the men invited to undergo PSA screening had a relative risk of death from prostate cancer of 0.79 (95% CI, 0.68 to 0.91). In eight randomized trials on screening for breast cancer,1 reductions in the relative risks of death from breast cancer were directly proportional to reductions in the relative risks of death from advanced breast cancer (slope, 1.0; 95% CI, 0.94 to 1.04), and 95% of the changes in risks of death were explained by changes in the risks of advanced cancer (R²=0.95, P<0.001). With the use of data from the Supplementary Appendix to the article by Schröder et al., we computed the relative risks of clinical stage T3 and T4 prostate cancer at each center participating in the ERSPC and plotted these risks against the risk of center-specific death from prostate cancer. There was no relationship between the relative risks of death and advanced cancer (slope, 0.09; 95% CI, 0.31 to 0.50), and changes in the risks of advanced cancer could not explain changes in risks of death (R²=0.039, P=0.62). This contrast suggests that factors other than PSA screening may be involved in the mortality reductions observed in the ERSPC.

Philippe Autier, M.D.
Mathieu Boniol, Ph.D.
International Prevention Research Institute, Lyon, France
philippe.autier@i-pri.org

Paul Perrin, M.D.
University Claude Bernard, Lyon, France

No potential conflict of interest relevant to this letter was reported.

1 Reference

1. 1

   Autier P, Hery C, Haukka J, Boniol M, Byrnes G. Advanced breast cancer and breast cancer mortality in randomized controlled trials on mammography screening. J Clin Oncol 2009;27:5919-5923
   CrossRef | Web of Science | Medline

To the Editor:

I was perplexed by Miller's editorial1 on the conflicting evidence regarding the efficacy of prostate-cancer screening. After astutely critiquing the two major randomized screening trials, one reporting no benefit (the PLCO trial2) and the other reporting a 21% reduction in prostate-cancer mortality (the ERSPC follow-up study), Miller concludes that physicians should no longer screen for prostate cancer. His stated rationale is that the U.S. study (the PLCO trial) was “more applicable to the situation in the United States.” However, the PLCO trial was clearly a study of more intensive screening versus less intensive screening, whereas the ERSPC examined screening versus no screening. Physicians are much more interested in the latter question: Should we screen at all? In fact, the mortality benefit reported in the ERSPC trial is similar to that of screening mammography.3 Clearly, the best available evidence remains divergent. When we are faced with such a dilemma, the reigning paradigm in American medicine is to involve the patient in an informed decision concerning whether to screen, not to abandon screening altogether.

Andrew M.D. Wolf, M.D.
University of Virginia School of Medicine, Charlottesville, VA

Dr. Wolf reports serving as chair of the American Cancer Society (ACS) Prostate Cancer Advisory Committee, which developed the 2010 ACS guideline for the early detection of prostate cancer. No other potential conflict of interest relevant to this letter was reported.

3 References

1. 1

   Miller AB. New data on prostate-cancer mortality after PSA screening. N Engl J Med 2012;366:1047-1048
   Full Text | Web of Science | Medline

2. 2

   Andriole GL, Crawford ED, Grubb RL III, et al. Prostate cancer screening in the randomized Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial: mortality results after 13 years of follow-up. J Natl Cancer Inst 2012;104:125-132
   CrossRef | Web of Science

3. 3

   Nelson HD, Tyne K, Naik A, et al. Screening for breast cancer: an update for the U.S. Preventive Services Task Force. Ann Intern Med 2009;151:727-737
   Web of Science | Medline

Author/Editor Response

In a departure from their recent meta-analysis,1 Djulbegovic et al., in their forest plot (see the Supplementary Appendix to their letter), use unadjusted outcomes that should not be directly compared. It is of questionable value to include small studies not designed to evaluate prostate-cancer mortality in randomized populations, such as the Stockholm trial (1782 men screened once) and the Norrköping trial (449 men screened once and 446 screened twice). The results of the Quebec trial are severely biased by a very low rate of participation in screening (23.6%) and by the use of different time frames for the time to death in the screening and control groups. The original design of the PLCO trial as a population-based screening study was confirmed in the revised calculation of its power by Prorok in 2000.2 However, because of contamination problems, the study ultimately became one that compared organized screening with nonorganized screening and therefore does not contribute to the evaluation of the effect of population-based screening.

For these reasons it is difficult to understand why the authors do not consider the effects of the original and follow-up ERSPC studies separately, since these studies were designed to evaluate the effect of population-based screening.

We fully agree with D'Amico and are grateful for his comparison of ERSPC and PLCO data, which points out important differences in the nature of the studies. His analysis promotes understanding, questions the meta-analysis by Djulbegovic et al., and helps clinicians in their discussions with men considering screening.

Autier et al. observed that rates of stage T3 and T4 prostate cancers do not differ between the screening and control groups of our study, whether considered according to center or overall. They conclude that mechanisms other than PSA screening could be involved in the mortality reduction observed in the ERSPC.

We see good reasons to contradict this observation. In contrast with breast cancer, in prostate cancer T stage alone is a poor predictor of outcome, as shown not only in our study but also in a large study of 4899 men who underwent radical prostatectomy.3 Other prognostic factors, such as metastatic stage, Gleason score, and the results of PSA screening, are probably more important. Direct comparisons between ERSPC centers are difficult owing to the short follow-up period and the small number of events. Other potential differences, such as differences in treatments, have not been found in the ERSPC overall or in Sweden.4,5

Fritz H. Schröder, M.D., Ph.D.
Erasmus University Medical Center, Rotterdam, the Netherlands
secr.schroder@erasmusmc.nl

Anssi Auvinen, M.D., Ph.D.
University of Tampere, Tampere, Finland

Jonas Hugosson, M.D., Ph.D.
Göteborg University, Göteborg, Sweden

Since publication of their article, the authors report no further potential conflict of interest.

5 References

1. 1

   Djulbegovic M, Beyth RJ, Neuberger MM, et al. Screening for prostate cancer: systematic review and meta-analysis of randomised controlled trials. BMJ 2010;341:c4543-c4543
   CrossRef | Web of Science

2. 2

   Prorok PC, Andriole GL, Bresalier RS, et al. Design of the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial. Control Clin Trials 2000;21:Suppl:273S-309S
   CrossRef | Medline

3. 3

   Reese AC, Cooperberg MR, Carroll PR. Minimal impact of clinical stage on prostate cancer prognosis among contemporary patients with clinically localized disease. J Urol 2010;184:114-119
   CrossRef | Web of Science

4. 4

   Wolters T, Roobol MJ, Steyerberg EW, et al. The effect of study arm on prostate cancer treatment in the large screening trial ERSPC. Int J Cancer 2010;126:2387-2393
   Web of Science | Medline

5. 5

   Hugosson J, Carlsson S, Aus G, et al. Mortality results from the Göteborg randomised population-based prostate-cancer screening trial. Lancet Oncol 2010;11:725-732
   CrossRef | Web of Science | Medline

Author/Editor Response

I am sorry that I failed to make it clear that I do not regard the evidence derived from the report of Schröder et al. as sufficient to justify screening for prostate cancer with the PSA test.1 The evidence on the harms of such screening in both the ERSPC and the PLCO trial1 is considerable. There was no benefit in superimposing organized screening on the practice at the time in the United States in the PLCO trial. In addition, the data so far reported from the ERSPC trial have been more compatible with the better treatment given to the screened participants as compared with controls than with any benefit derived from the earlier detection of cancer as determined by the limited amount of PSA screening in that trial.1

Physicians were largely responsible for purveying the information to their patients that PSA screening was beneficial. It is now the responsibility of physicians to correct that error, recognizing that the interaction with patients needs to emphasize the lack of balance in the evidence, a position that the ERSPC investigators endorse.

Anthony B. Miller, M.D.
University of Toronto, Toronto, ON, Canada

Since publication of his article, the author reports no further potential conflict of interest.

1 Reference

1. 1

   Andriole GL, Crawford ED, Grubb RL III, et al. Prostate cancer screening in the randomized Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial: mortality results after 13 years of follow-up. J Natl Cancer Inst 2012;104:125-132
   CrossRef | Web of Science