Evidence-Based Medicine Online First, published on June 17, 2014 as 10.1136/ebmed-2014-110018 Prognosis Cohort study
A targeted approach reduces prostate cancer-specific (PSA) screening harms while preserving benefits 10.1136/ebmed-2014-110018
Amanda Black Division of Cancer Epidemiology & Genetics, National Cancer Institute, Rockville, Maryland, USA Correspondence to: Dr Amanda Black, Division of Cancer Epidemiology & Genetics, National Cancer Institute, 9609 Medical Center Drive, Room 7E-582, Rockville, MD 20892, USA; [email protected]
Commentary on: Carlsson S, Assel M, Sjoberg D, et al. Influence of blood prostate specific antigen levels at age 60 on benefits and harms of prostate cancer screening: population based cohort study. BMJ 2014;348:g2296.
Context Prostate-specific antigen (PSA) screening for prostate cancer may prevent some men dying from the disease.1 However, had they not been screened, many contemporary men diagnosed with prostate cancer would never have experienced symptoms or died from their cancer. Considering the devastating impact of an unnecessary cancer diagnosis and treatment-associated sequelae, PSA screening as it is currently used arguably results in greater harm than good. Nevertheless, PSA is entrenched in healthcare systems across the globe and, until we identify another test that can accurately discriminate the few potentially fatal malignancies from the many more indolent prostate cancers, PSA screening will likely remain commonplace. As such, strategies to improve the use of PSA screening are critically important to limit harms.
Methods Although described as a population-based cohort, this study is actually a comparison of a sample of men from each of two population-based cohorts in Sweden, one composed of participants in the intervention (screened) arm of the Gothenburg site of the European Randomized Screening for Prostate Cancer trial (ERSPC) and the other composed of individuals from the earlier Malmo Preventive Project (unscreened). Included were 1162 men from Malmo, aged 60 years at blood draw and 1756 men from Gothenburg aged 57.5–62.5 years at blood draw. PSA was directly quantitated for all men in the Gothenburg group and for half of the Malmo group (ie, all of the cases and three matched controls). PSA values were imputed for the remaining Malmo participants. The effect of screening on prostate cancer incidence, metastatic disease and prostate cancer-specific mortality, through 15 years of follow-up, was assessed using cumulative incidence rates and risk differences stratified by baseline PSA level at age 60 (or close to 60). The number of men needed to be screened (NNS) and the number of men needed to be diagnosed (NND) in order to prevent one metastatic cancer or death from prostate cancer within 15 years was also reported.
Findings In the Gothenburg group, 318 (18%) men were diagnosed with prostate cancer, 22 (1.3%) experienced metastatic disease, and 14 (0.8%) died from prostate cancer during the 15-year follow-up period. The corresponding numbers in Malmo were 63 (5.4%) prostate cancers, 20 (1.7%) metastatic cases and 16 (1.4%) deaths. Overall, there was no difference in risk of prostate cancer-specific mortality between the two groups. In the Malmo group, there were zero events of metastatic disease in men with a PSA of
A targeted approach reduces prostate cancer-specific (PSA) screening harms while preserving benefits 10.1136/ebmed-2014-110018
Amanda Black Division of Cancer Epidemiology & Genetics, National Cancer Institute, Rockville, Maryland, USA Correspondence to: Dr Amanda Black, Division of Cancer Epidemiology & Genetics, National Cancer Institute, 9609 Medical Center Drive, Room 7E-582, Rockville, MD 20892, USA; [email protected]
Commentary on: Carlsson S, Assel M, Sjoberg D, et al. Influence of blood prostate specific antigen levels at age 60 on benefits and harms of prostate cancer screening: population based cohort study. BMJ 2014;348:g2296.
Context Prostate-specific antigen (PSA) screening for prostate cancer may prevent some men dying from the disease.1 However, had they not been screened, many contemporary men diagnosed with prostate cancer would never have experienced symptoms or died from their cancer. Considering the devastating impact of an unnecessary cancer diagnosis and treatment-associated sequelae, PSA screening as it is currently used arguably results in greater harm than good. Nevertheless, PSA is entrenched in healthcare systems across the globe and, until we identify another test that can accurately discriminate the few potentially fatal malignancies from the many more indolent prostate cancers, PSA screening will likely remain commonplace. As such, strategies to improve the use of PSA screening are critically important to limit harms.
Methods Although described as a population-based cohort, this study is actually a comparison of a sample of men from each of two population-based cohorts in Sweden, one composed of participants in the intervention (screened) arm of the Gothenburg site of the European Randomized Screening for Prostate Cancer trial (ERSPC) and the other composed of individuals from the earlier Malmo Preventive Project (unscreened). Included were 1162 men from Malmo, aged 60 years at blood draw and 1756 men from Gothenburg aged 57.5–62.5 years at blood draw. PSA was directly quantitated for all men in the Gothenburg group and for half of the Malmo group (ie, all of the cases and three matched controls). PSA values were imputed for the remaining Malmo participants. The effect of screening on prostate cancer incidence, metastatic disease and prostate cancer-specific mortality, through 15 years of follow-up, was assessed using cumulative incidence rates and risk differences stratified by baseline PSA level at age 60 (or close to 60). The number of men needed to be screened (NNS) and the number of men needed to be diagnosed (NND) in order to prevent one metastatic cancer or death from prostate cancer within 15 years was also reported.
Findings In the Gothenburg group, 318 (18%) men were diagnosed with prostate cancer, 22 (1.3%) experienced metastatic disease, and 14 (0.8%) died from prostate cancer during the 15-year follow-up period. The corresponding numbers in Malmo were 63 (5.4%) prostate cancers, 20 (1.7%) metastatic cases and 16 (1.4%) deaths. Overall, there was no difference in risk of prostate cancer-specific mortality between the two groups. In the Malmo group, there were zero events of metastatic disease in men with a PSA of
