Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
QUESTION: How effective is breast cancer screening with mammography, clinical breast examination (CBE), and breast self examination (BSE) in preventing breast cancer mortality?
Studies were identified by searching Medline (1994–2001), Premedline (December 2001 to February 2002), and the Cochrane Controlled Trials Register; reviewing the reference lists of previous reviews, commentaries, and meta-analyses; and contacting experts in the field.
Studies were selected if they were randomised controlled trials (RCTs) of breast cancer screening and had a relevant clinical outcome (advanced breast cancer, breast cancer mortality, or all cause mortality).
Data were extracted on patient population, study design, potential flaws, missing information, analysis, and length of follow up. US Preventive Services Task Force (USPSTF) criteria were used to assess study quality (good, fair, or poor). The primary endpoint was breast cancer mortality.
8 RCTs (479 987 women) (154 publications) met the selection criteria: 4 evaluated mammography, and 4 evaluated mammography plus CBE. 7 trials were rated fair quality, and 1 was rated poor quality. The mean follow up was 14 years. Meta-analysis of the 7 fair trials showed that mammography screening led to a modest reduction in breast cancer mortality across all age groups, with greater benefit conferred in older women (table). The results were consistent with those of 5 of 7 previous meta-analyses identified in the search for trials. Mortality reductions in trials of mammography plus CBE were similar to those of trials of mammography alone. 2 RCTs of BSE met the selection criteria. Both trials showed no difference in breast cancer mortality rates in women instructed in BSE and in non-instructed women.
Fair quality evidence from randomised controlled trials shows that mammography screening is effective in reducing breast cancer mortality. Biennial and annual screening are equally effective. Clinical breast examination confers no additional benefit. Instruction in breast self examination, as a single screening method, is ineffective.
At first glance, these 2 reports seem inconsistent: The study by Miller et al is an update of the Canadian National Breast Screening Study, continuing to show no hint of benefit in women 40–49 years of age; the review by Humphrey et al is a USPSTF meta-analysis extending a recommendation for mammography to this age group for the first time. Ironically, the Canadian study was judged to be the highest quality study of all those evaluated by the USPSTF. Even with >50 000 participants, the Canadian study did not have the power to detect a protective effect <40%; the USPSTF meta-analysis found a much smaller benefit of 15% in the 40–49 year age group. The 95% confidence intervals/credible intervals of the reduction in breast cancer mortality in the 2 studies (0.74 to 1.27 v 0.73 to 0.99) overlap widely. The absolute reduction in breast cancer mortality is low in all age groups. For women 40–49 years of age, it is estimated to be <1/10 000 per year. For older women, the benefit is slightly greater and the confidence intervals more clearly exclude the null result.
At what price is this modest benefit obtained? Surprisingly, in the appendix the USPSTF states: “A systematic review of adverse effects was beyond the scope of our review”. The evidence for adverse effects includes false positive results, believed to occur in 6.5% of mammograms;1 radiogenic breast cancer, estimated by the USPSTF to negate <10% of the breast mortality benefit; and diagnosis and treatment of cancer that may not become clinically important. In the Canadian study, about 14% more diagnoses of in situ or invasive breast cancer occurred in the screened group than in the usual care group; this difference persisted to almost 10 years after the intervention ended. In 2 Swedish trials including older women, 35% more women received major surgery (mastectomy or lumpectomy) and 25% more women received radiation in the screened groups.2 These results suggest that this may be a more substantial problem than previously thought.
How can we convey these results to our patients? For women 40–49 years of age, the estimated benefit of mammography is small (15%, or <1/10 000 breast cancer deaths prevented per year) and the evidence of benefit is weak, with confidence intervals nearly overlapping 1. The risk of false positive results is higher at this age, as is the potential for radiation carcinogenesis.3 The absolute benefit might be greater in women at high risk because of a positive family history. However, mammography has been shown to be less sensitive in this group,4 and if the family history results from inherited radiation sensitivity,5 the risk of radiogenic breast cancer will be further increased.
For women ≥50 years, the evidence of a 20–25% relative benefit in breast cancer mortality is stronger and exceeds 1/10 000 per year. An important issue in this group is the discovery of cancer that might not have caused symptoms, especially in women with comorbid conditions and a limited life span.
Unfortunately, it is clear that most breast cancer deaths will not be prevented by mammography at any age. Perhaps the enormous resources devoted to the debate, promotion, and provision of mammography could be better used to study the efficacy of more sensitive detection systems6 or to develop predictive models with greater discriminatory power.7
For correspondence: Dr L L Humphrey, Oregon Health & Science University, Portland, OR, USA.
Abstract and commentary also appear in ACP Journal Club.
Source of funding: Agency for Healthcare Research and Quality.