Mammography screening increases non-breast cancer mortality
Peter Gøtzsche, Nordic Cochrane Centre
5 December 2013
Erpeldinger et al. used the randomised trials to find out whether mammography screening increases non-breast cancer mortality. As we have explained in our Cochrane review (1), to which the authors refer, this cannot be done. One of the problems is that assignment of cause of death was biased in these trials, but the most important problem we describe in our review is this one:
"The complex designs and insufficient reporting precluded us from providing reliable estimates for all-cause mortality in the trials with suboptimal randomisation. Furthermore, these trials had introduced early screening of the control group or had differentially excluded women after
randomisation (1)."
Erpeldinger et al. used data after 13 years for their calculations, which is clearly inappropriate when the whole control group has been screened long before this cut-off, which was the case in all the Swedish trials apart from the Malmö trial. Screening the control group increases non-breast cancer mortality in this group, which means it gets close to impossible to find an effect of screening on non-breast cancer mortality in a meta-analysis of the randomised trials.
Screening leads to huge overdiagnosis of healthy women, about 30% in the trials (1) and about 50% in countries with publicly organised screening programmes (2). Many of these women are treated with radiotherapy, which can only be harmful when given to healthy overdiagnosed women. Mike Baum has estimated that radiotherapy of overdiagnosed women kills at least as many as screening saves from dying from breast cancer (3), even assuming that screening reduces breast cancer mortality by 20%, which it cannot do; tumour data suggest that the effect in the old trials, before effective treatments came into common usage, cannot have exceeded 12% (4). This means that screening has no effect at all on mortality whereas it leads to tremendous psychosocial harm because of the false positive findings (5), which 10% of the women will have experienced already after 10 years of screening. Screening is harmful and should be stopped (6).
1 Gøtzsche PC, Jørgensen KJ. Screening for breast cancer with mammography. Cochrane Database of Systematic Reviews 2013, Issue 6. Art. No.: CD001877.
3 Baum M. Harms from breast cancer screening outweigh benefits if death caused by treatment
is included. BMJ 2013; 346: f385.
4 Gøtzsche PC, Jørgensen KJ, Zahl PH, Maehlen J. Why mammography screening has not lived up to expectations from the randomised trials. Cancer Causes Control 2012; 23(1): 15-21.
5 Brodersen J, Siersma V. Long-term psychosocial consequences of false-positive screening mammography - a cohort study with 3-year follow-up. Annals of Family Medicine 2013;11(2):106–15.
6 Gøtzsche PC. Re: Time to stop mammography screening? CMAJ 2011; (183): 1957-8.
Trials
Mammography screening increases non-breast cancer mortality
5 December 2013
Erpeldinger et al. used the randomised trials to find out whether mammography screening increases non-breast cancer mortality. As we have explained in our Cochrane review (1), to which the authors refer, this cannot be done. One of the problems is that assignment of cause of death was biased in these trials, but the most important problem we describe in our review is this one:
"The complex designs and insufficient reporting precluded us from providing reliable estimates for all-cause mortality in the trials with suboptimal randomisation. Furthermore, these trials had introduced early screening of the control group or had differentially excluded women after
randomisation (1)."
Erpeldinger et al. used data after 13 years for their calculations, which is clearly inappropriate when the whole control group has been screened long before this cut-off, which was the case in all the Swedish trials apart from the Malmö trial. Screening the control group increases non-breast cancer mortality in this group, which means it gets close to impossible to find an effect of screening on non-breast cancer mortality in a meta-analysis of the randomised trials.
Screening leads to huge overdiagnosis of healthy women, about 30% in the trials (1) and about 50% in countries with publicly organised screening programmes (2). Many of these women are treated with radiotherapy, which can only be harmful when given to healthy overdiagnosed women. Mike Baum has estimated that radiotherapy of overdiagnosed women kills at least as many as screening saves from dying from breast cancer (3), even assuming that screening reduces breast cancer mortality by 20%, which it cannot do; tumour data suggest that the effect in the old trials, before effective treatments came into common usage, cannot have exceeded 12% (4). This means that screening has no effect at all on mortality whereas it leads to tremendous psychosocial harm because of the false positive findings (5), which 10% of the women will have experienced already after 10 years of screening. Screening is harmful and should be stopped (6).
1 Gøtzsche PC, Jørgensen KJ. Screening for breast cancer with mammography. Cochrane Database of Systematic Reviews 2013, Issue 6. Art. No.: CD001877.
2 Jørgensen KJ, Gøtzsche PC. Overdiagnosis in publicly organised mammography screening programmes: systematic review of incidence trends. BMJ 2009; 339: b2587.
3 Baum M. Harms from breast cancer screening outweigh benefits if death caused by treatment
is included. BMJ 2013; 346: f385.
4 Gøtzsche PC, Jørgensen KJ, Zahl PH, Maehlen J. Why mammography screening has not lived up to expectations from the randomised trials. Cancer Causes Control 2012; 23(1): 15-21.
5 Brodersen J, Siersma V. Long-term psychosocial consequences of false-positive screening mammography - a cohort study with 3-year follow-up. Annals of Family Medicine 2013;11(2):106–15.
6 Gøtzsche PC. Re: Time to stop mammography screening? CMAJ 2011; (183): 1957-8.
Competing interests
None.