Mammography reduces long-term mortality for women aged 50 to 74 years
ACP J Club. 1995 Sept-Oct;123:34. doi:10.7326/ACPJC-1995-123-2-034
Kerlikowske K, Grady D, Rubin SM, Sandrock C, Ernster VL. Efficacy of screening mammography. A meta-analysis. JAMA. Jan
To determine the efficacy of screening mammography by age category, number of views per screening, screening interval, and length of follow-up.
English-language studies were identified through MEDLINE (1966 to October 1993) using the terms breast neoplasm, mortality, mass screening, and female; from bibliographies of relevant studies; and through contact with experts and colleagues.
Studies were selected if they were randomized controlled trials (RCTs) or cohort or case-control studies that evaluated the efficacy of screening mammography; if the main outcome was mortality; if > 10 deaths from breast cancer occurred; if follow-up was ≥ 5 years; if relative risk (RR) or odds ratios (OR) were reported with CIs; and if data were adjusted for age or age-matched women were used as controls. Hospital-based studies and uncontrolled cohorts were excluded. 1 unpublished study and 15 reports on 13 studies (9 RCTs and 4 case-control studies) were selected from 264 articles.
Age, start date and type of study, screening interval, number of mammographic views, use of clinical breast examination, length of screening, length of follow-up, and RR or OR for mortality.
Case-control studies had a lower summary RR for mortality than RCTs (0.62 vs 0.79) and were excluded from all further analyses. Studies begun before 1980 had a lower RR than studies started later (0.77 vs. 0.87). The RRs for older women (50 to 74 y) did not differ for comparison of 1- and 2-view mammography (0.70 vs. 0.83); mammography done every 12 months compared with every 18 to 33 months (0.77 for each); 7 to 9 years of follow-up compared with 10 to 12 years of follow-up (0.73 vs. 0.76); screening duration of 3 to 5 years compared with 8 to 10 years (0.76 vs. 0.78); and the addition of clinical breast examination compared with mammography alone (0.76 vs. 0.78). For women aged 50 to 74 years at screening, the overall summary RR estimate from RCTs was 0.77 (95% CI, 0.69 to 0.87). For women aged 40 to 49 years, mammography did not decrease mortality.
For women aged 50 to 74 years, screening mammography reduces mortality from breast cancer after 7 to 9 years of follow-up. This reduction is present regardless of screening interval, number of mammography views used, length of follow-up, or the addition of clinical breast examination. For women aged 40 to 49 years, screening mammography does not reduce mortality even after 10 to 12 years of follow-up.
Source of funding: National Institutes of Health.
For article reprint: Dr. K. Kerlikowske, San Francisco Veterans Affairs Medical Center, General Internal Medicine Section, 111A1, 4150 Clement Street, San Francisco, CA 94121. FAX 415-386-4044.
The article by Kerlikowske and colleagues provides important information that should help resolve the current controversy over the role of screening mammography in women < 50 years of age. This elegant meta-analysis clearly shows a survival benefit for screening mammography in women between the ages of 50 and 74 years. A 26% reduction in mortality was identified in this group of women after 7 to 9 years of follow-up, regardless of screening interval or number of mammographic views (single vs. double). In contrast, no evidence of a survival benefit was seen among women between the ages of 40 and 49 years; an overall nonsignificant risk reduction of 7% was identified. A nonsignificant increase in risk (2%) was identified at 7 to 9 years of follow-up, and a nonsignificant reduction in risk of 17% appeared after 10 to 12 years of follow-up. The authors suggested that the latter finding reflected the fact that younger women became postmenopausal with prolonged surveillance and that the late benefits were a result of screening women while they were postmenopausal or more than 50 years old and not a result of screening women in their early forties.
Although reducing breast cancer mortality among younger women is highly desirable, the results of this meta-analysis clearly show that the routine use of mammographic screening in women aged 40 to 49 years will not achieve this goal. As was so clearly pointed out by the authors, the issue is not low statistical power, it is simply that mammographic screening does not reduce breast cancer mortality in this group of women. Continued recommendations that population-based mammographic screening should be done in younger women cannot be justified by the existing evidence. The health care resources currently used for this activity would probably have a greater effect on breast cancer mortality if they were allocated to the development of new screening techniques for these younger women; to the evaluation of preventive strategies, such as anti-estrogen and dietary fat reduction; or to the development of new treatment strategies.
The article by Kattlove and colleagues examined the costs and benefits of screening and treatment for early breast cancer from the perspective of a hypothetical health care organization of 500 000 persons. Using data available in the literature, they modeled the beneficial effects of screening mammography in older women, adjuvant chemotherapy in premenopausal women, and adjuvant hormonochemotherapy in postmenopausal women. Using conservative approaches to costing that considered only the "up-front" costs of these treatments and not the costs saved as a result of prevention of metastatic breast cancer, they clearly show acceptable costs associated with mammographic screening in women aged 50 to 69 years and with some forms of adjuvant drug treatment. Had the investigators considered the treatment of metastatic disease, the resulting costs would be even more acceptable. The investigators have also shown a policy of offering a choice of primary surgical treatment (lumpectomy with radiation vs. mastectomy) to be associated with acceptable cost.
In contrast, 2 approaches to management—mammographic screening in women aged < 50 years and routine investigation for metastatic disease in patients who have completed primary treatment—have not been shown to improve clinical outcomes and could not be justified from a cost perspective. This observation is important because it shows the negative cost effect of 2 common interventions that are often considered justified because they "do no harm." The harm may come from the drain of health care dollars away from interventions that do improve outcomes. The costing data provide additional justification, if any is really needed, for not offering these interventions as part of routine clinical practice.
A common message in these 2 articles is that sometimes it is best to do nothing. This may be a difficult message for compassionate, patient-oriented clinicians to accept, particularly when dealing with a devastating illness like breast cancer, but these articles suggest that the time may have come to do just that.
Pamela J. Goodwin, MD, MSc
Mount Sinai Hospital Toronto, Ontario