Esserman and Kerlikowske have done an excellent job in reviewing the factual information on screening mammography for women age 40 to 49 years. Their review builds on some previously published work by Kerlikowske and colleagues, particularly their meta-analysis . This meta-analysis was important, in that it addressed the issue of timing in relation to mammography screening in women 40 to 49 years old, as compared with those 50 to 69 years of age. The combined data of eight randomized trials clearly demonstrated that there was absolutely no benefit of mammography for women age 40 to 49 at 7 to 9 years after the initiation of screening. In contrast, for women age 50 to 69, there was a substantial and statistically significant reduction in breast cancer mortality.
Esserman and Kerlikowske have done an excellent job in reviewingthe factual information on screening mammography for women age40 to 49 years. Their review builds on some previously publishedwork by Kerlikowske and colleagues, particularly their meta-analysis. This meta-analysis was important, in that it addressed theissue of timing in relation to mammography screening in women40 to 49 years old, as compared with those 50 to 69 years of age.The combined data of eight randomized trials clearly demonstratedthat there was absolutely no benefit of mammography for womenage 40 to 49 at 7 to 9 years after the initiation of screening.In contrast, for women age 50 to 69, there was a substantial andstatistically significant reduction in breast cancer mortality.
Over the period of 10 to 12 years from the initiation of screening,there was some suggestion of a nonsignificant reduction in mortalityfor women 40 to 49 years old, but this was nowhere near as greatas the persistence of a similar order of benefit at 10 to 12 yearsas for 7 to 9 years for women age 50 to 69. The meta-analysistherefore demonstrated the continued validity of a conclusionwe reached 5 years before; namely, that there is no evidence ofbenefit from screening women in their 40s, at least in the first10 years after the initiation of screening .
Delayed Benefit Seems Biologically Implausible
Although several researchers continue to dispute the lack of benefitin the younger age group and suggest that it is not unreasonablefor a delayed benefit to occur from screening, they are un ableto convince us of the validity of these beliefs, largely becausethere does not seem to be a biologic reason why there should bea greater delay for seeing benefit from screening in younger comparedto older women. In addition, much of the data have accrued fromwomen who commenced screening at age 45 to 49, and a delayed benefitin relation to the initiation of screening at these ages is entirelycompatible with the anticipated effect of screening when womenreach the age of 50. This point is well considered by Essermanand Kerlikowske, who refer to the important work by De Koninget al  in modeling expected benefits.
Esserman and Kerlikowske point to a possible reason for the nonsignificantexcess breast cancer mortality we observed in the earlier stagesof the Canadian National Breast Screening Study (CNBSS), suggestingthis may have been due to a delay in diagnosis. We have carefullyevaluated our data, and find no evidence of such a delay. Oneof the reasons for this may have been our teaching and subsequentreinforcement of breast self-examination, a procedure which, onits own, now appears to be providing some benefit to both youngerand older women .
Women With a Positive Family History
Esserman and Kerlikowske also discuss the lack of informationon the benefit of screening mammography among women with a positivefamily history. I would add additional caution to their remarks.There may be some risk in using screening to attempt to providesupport for such women. There is no reason to believe that womenwith a positive family history, particularly those carrying breastcancer susceptibility genes, would benefit from screening whenwomen at normal risk do not. Indeed, some have expressed concernthat there may be detriments from the use of mammography in womenat increased risk--a possibility that we will be evaluating overthe next few years using the data on family history accumulatedin the CNBSS supplemented by additional data collected from theparticipants or their relatives.
Physical Exams in Women Aged 40 to 49
In debates over this issue, the point is often made that if mammographyis not beneficial in women aged 40 to 49, the trials suggest thatphysical examination is not either. In practice, the trial evidencecannot be used to draw such a conclusion. One of the trials thatsuggests a delayed benefit, the HIP trial, used good physicalexamination but 1960s mammography, and only a fraction of thecancers were diagnosed by mammography alone. None of the Swedishtrials used physical examination. In the Edinburgh trial, whichsuggests a benefit from screening in this age group, physicalexaminations were performed every year and mammography every 2years, and yet women were only recruited from the age of 45; therefore,this trial cannot fully evaluate the question.
In the CNBSS, physical examination was given to both groups uponrecruitment, and women in the control arm were urged to continueto practice breast self-examination and were reminded annuallyto do so and to see their own doctor for a physical examination.Thus, the level of physical examination screening in the controlgroup of this trial was probably substantially greater than thatin normal practice. What we demonstrated in this younger age group,as we did in the older, is that mammography does not produce amortality benefit, over and above the physical examination andthe practice of self breast-examination. However, we cannot addressdirectly the effect of physical examination.
Additional Costs of Screening
As if the lack of benefit of screen-ing mammography is not enough,Esserman and Kerlikowske point to the additional costs from detectingductal carcinoma in situ in this age group, as well as the additionalbiopsies. I would only add that not only were we able to showsuch adverse effects, and thus, excess health-care costs, butalso we found no reduction in the rate of mastectomies in themammography arm .
Thus, at present the appropriate recommendation for women andtheir physicians is to practice breast self- examination, havean annual physical examination by a physician or other healthprofessional who knows the signs of early breast cancer, and usemammography if indicated by either of these procedures as a diagnostictool, but not for routine screening.
1. Kerlikowske K, Grady D, Rubin SM et al: Efficacy of screeningmammography: A meta-analysis. JAMA 273:149-154, 1995.
2. Miller AB, Chamberlain J, Day NE et al: Report on a workshopof the UICC Project on evaluation of screening for cancer. IntJ Cancer 46:761-769, 1990.
3. De Koning HJ, Boer R, Wannerdain PG et al: Interpretation ofage-specific mortality reductions from the Swedish breast cancerscreening trials. J Natl Cancer Inst 87:1217-1223, 1995.
4. Harvey BJ, Miller AB, Baines CJ, et al: A case-control studyof breast self-examination practice nested in the Canadian NationalBreast Screening Study. Submitted to NEJM.
5. Miller AB: May we agree to disagree, or how do we develop guidelinesfor breast cancer screening in women? J Natl Cancer Inst 86:1729-1731,1994.