Correspondence
Breast-Feeding and Breast Cancer
N Engl J Med 1994; 330:1682-1684June 9, 1994
- Article
To the Editor:
In their interesting report, Newcomb et al. (Jan. 13 issue)1 do not speculate about why the reduced risk of breast cancer associated with lactation was limited to premenopausal women. Changes in breast-feeding practices in the United States over recent decades may help explain this finding and may shed light on why no association with lactation was observed among premenopausal women in the Nurses' Health Study2.
In the first half of this century, a rigid schedule of five breast-feedings per day was widely recommended. Dr. Spock's popular manual suggested greater flexibility in responding to an infant's cries of hunger but still held out the ideal of five feedings per day within one to two months of birth3. More recent advice for lactating mothers emphasizes feeding “on demand” -- whenever the infant seems hungry -- which may often be eight or more times per day during the first several months4.
Among women with an equal duration of lactation, less frequent breast-feeding promotes the earlier return of ovulation5. Whether a putative protective effect of lactation is mediated by the suppression of ovulation or by direct physical changes in the breast,1 women who nurse their babies on demand would be expected to be protected more than those who followed earlier recommendations to nurse less frequently.
In the case of premenopausal women in the study by Newcomb et al.,1 who were approximately 30 to 50 years old in 1989, the majority of their lactation occurred during the 1970s and 1980s, when demand feeding became popular. In contrast, the postmenopausal women in that study and the women in the Nurses' Health Study2 did most of their breast-feeding in earlier decades, when a limited feeding schedule was recommended. Thus, on average, premenopausal women in the study of Newcomb et al. probably nursed more frequently for a given duration of lactation than either postmenopausal women in the same study or women in the Nurses' Health Study.
If lactation truly decreases the risk of breast cancer, this change in lactation practices may help explain why the association observed by Newcomb et al. between lactation and breast cancer was limited to premenopausal women and why such an association was not observed in the Nurses' Health Study or some earlier U.S. studies of lactation and breast cancer cited by Newcomb et al.
5 ReferencesStephanie J. London, M.D., Dr.P.H.
University of Southern California School of Medicine, Los Angeles, CA 900331
Newcomb PA ,Storer BE ,Longnecker MP , et al. Lactation and a reduced risk of premenopausal breast cancer. N Engl J Med 1994;330:81-87
Full Text | Web of Science | Medline2
London SJ ,Colditz GA ,Stampfer MJ , et al. Lactation and risk of breast cancer in a cohort of US women. Am J Epidemiol 1990;132:17-26
Web of Science | Medline3
Spock B. The common sense book of baby and child care. Rev. ed. New York: Duell, Sloan and Pearce, 1957.
4
Eiger MS, Olds SW. The complete book of breast feeding. Rev. ed. New York: Bantam Books, 1987.
5
Gray RH ,Campbell OM ,Apelo R , et al. Risk of ovulation during lactation. Lancet 1990;335:25-29
CrossRef | Web of Science | Medline
To the Editor:
The article by Newcomb et al. was a case-control study by design. However, the authors chose to use the relative risk instead of the odds ratio to measure the risk of breast cancer in relation to lactation. The authors interpreted the results as reduction in the risk of breast cancer that was caused by lactation, instead of describing lactation as associated with a reduced occurrence of breast cancer. The appropriate result should be expressed thus: “After adjustment for other factors, the odds of breast cancer in premenopausal women who had lactated was 0.78.”
The importance of this interpretation lies in the fact that there is an association between lactation and a reduced occurrence of breast cancer among premenopausal women. The authors' interpretation implies that if you lactated, your risk of premenopausal breast cancer is reduced, and hence this should be a routine recommendation. I am not certain how this slipped by the editorial board and statisticians.
Anand B. Mutgi, M.D.
Medical College of Ohio, Toledo, OH 43699-0008To the Editor:
The paucity of detail Newcomb et al. provide about the nursing habits of the women they studied raises questions about the validity of their results. A mother who lactates for three months might nurse on demand 24 hours a day; nurse morning and night, using formula during the day while she works; or pump her breasts twice a day while supplementing with first 4 oz (120 ml), then 8 oz (235 ml), then 12 oz (350 ml) of formula per day.
All these women might legitimately describe themselves as lactating, and all of these nursing styles can be voluntary, rather than forced by insufficient milk supply. But the effect on the breast tissue, and thus on the risk of breast cancer, is certainly not equivalent. Grouping such women together is akin to assigning the risk of lung cancer according to years of smoking, without regard to the number of packs smoked per day, or the risk of cirrhosis according to years of drinking, without regard to the amount of alcohol consumed.
I await the publication of a study that divides lactating women into more homogeneous groups in order to provide a more definitive estimate of the reduction in the risk of breast cancer actually associated with lactation.
Diane Lind Fenster, M.D.
1625 Patton St., Green Bay, WI 54301To the Editor:
Newcomb et al. report that lactation reduces the risk of breast cancer, that this protection is confined to premenopausal women, and that it is magnified by an early age at first lactation. They report that previous studies of this relation have shown only small, inconsistent effects. We have published the results of four population-based case-control studies of breast cancer conducted in the three largest urban areas in China -- Shanghai,1 Beijing,2 and Tianjin3,4.
All four studies found protective associations that varied with the duration of lactation. Despite the fact that in each of these studies every major risk factor for breast cancer in Western populations was clearly identifiable (suggesting that the pathogenesis of breast cancer is similar in Chinese women and in women in Western societies), the effects of lactation were unrelated to menopausal status (Table 1Table 1
Duration of Lactation among Parous Women with Breast Cancer and Controls in China.). Because we can think of no mechanism for an effect of lactation on the pathogenesis of breast cancer that could create a distinct reduction in risk that is totally confined to the premenopausal period, our results are what we would have anticipated.Though we ourselves have emphasized the importance of early life events as critical determinants of the risk of breast cancer,5 we believe that it is difficult to tease out the effects of age at first lactation (or, for that matter, duration of lactation) from the highly correlated risk factors of parity and age at first full-term pregnancy. In fact, since nearly every woman would be expected to have her first experience of lactation immediately after her first full-term pregnancy, age at first delivery and age at first lactation would become, in effect, the same variable (in the Chinese data, this was true for 95 percent of the case subjects and 96 percent of the controls).
We are surprised that Newcomb et al. did not find age at first delivery to be a confounder of age at first lactation. We are also surprised that they report essentially no change in the relative risk associated with lactation after adjustment for parity and age at first delivery. Adjustment for these two variables had a definite effect on the association between lactation and breast cancer among the Chinese women we studied.
Another potential source of confounding is socioeconomic status, which was not adjusted for in the analysis by Newcomb et al. Among the Chinese women in our study, adjustment for the level of education (as a surrogate for socioeconomic status) further reduced the effect of lactation on the risk of breast cancer.
5 ReferencesRonald K. Ross, M.D.
Mimi C. Yu, Ph.D.
Norris Comprehensive Cancer Center, Los Angeles, CA 90033-08041
Yuan J-M ,Yu MC ,Ross RK ,Gao Y-T ,Henderson BE . Risk factors for breast cancer in Chinese women in Shanghai. Cancer Res 1988;48:1949-1953
Web of Science | Medline2
Tao S-C ,Yu MC ,Ross RK ,Xiu K-W . Risk factors for breast cancer in Chinese women of Beijing. Int J Cancer 1988;42:495-498
CrossRef | Web of Science | Medline3
Wang Q-S, Yu MC, Henderson BE. Risk factors for breast cancer in Tianjin, People's Republic of China. In: Fourth Symposium on Epidemiology and Cancer Registries in the Pacific Basin. National Cancer Institute Monograph 69. Washington, D.C.: Government Printing Office, 1985:39-42. (NIH publication no. 85-2768).
4
Wang Q-S ,Ross RK ,Yu MC ,Ning J-P ,Henderson BE ,Kimm HT . A case-control study of breast cancer in Tianjin, China. Cancer Epidemiol Biomarkers Prev 1992;1:435-439
Web of Science | Medline5
Henderson BE ,Ross RK ,Pike MC . Hormonal chemoprevention of cancer in women. Science 1993;259:633-638
CrossRef | Web of Science | Medline
Author/Editor Response
The authors reply:
To the Editor: Ross and Yu's findings among Chinese women are unusual in their consistency among both premenopausal and postmenopausal women. Differences between the populations in their study and ours are notable. In particular, the proportion of premenopausal women who had borne children but had never lactated in their study was about 8 percent, as compared with about 40 percent in our study. In their study, therefore, age at first lactation and age at first delivery would be more highly correlated than in our study, in which we had greater opportunity to assess the contribution of each factor.
Another striking difference is the proportion of cases of breast cancer in premenopausal women -- about 70 percent of the total case group -- in their study. In our study, cases in premenopausal women made up only 23 percent of the total. Although Ross and Yu claim that every major Western risk factor was observed in their study, the large proportion of cases in premenopausal women is remarkable and suggests dissimilar causes of cancer in the two populations. In our relatively homogeneous study population, education was not a confounder; the adjusted odds ratio, including the level of education, for premenopausal women who had ever lactated was 0.78, similar to the estimate without the adjustment for education.
Both London and Fenster note that breast-feeding practices vary and that current behavior may reflect more frequent on-demand feeding. Although the specific mechanisms whereby lactation may reduce the risk of breast cancer are not fully understood, less frequent nursing and earlier return of menses probably affect the relation. Thus, the results of our study may differ from those in other populations, such as the women in the Nurses' Health Study, in which the pattern of lactation was different. Furthermore, many of the studies that show an association with premenopausal breast cancer were conducted many years ago1. Other risk factors for breast cancer -- such as weight2,3 and perhaps other reproductive variables3 -- appear to be modified by menopausal status.
We share Mutgi's concern about the correct interpretation of case-control data; however, the assertion that the term “relative risk” cannot be used in a generic sense to refer to the odds ratio is simply incorrect4. In fact, for these data, the relative risk and the odds ratio are essentially equivalent; neither term implies a cause-and-effect relation.
4 ReferencesPolly A. Newcomb, Ph.D.
Barry E. Storer, Ph.D.
University of Wisconsin, Madison, WI 53706Matthew P. Longnecker, M.D., Sc.D.
UCLA School of Public Health, Los Angeles, CA 900241
Byers TM ,Graham S ,Rzepka T ,Marshall J . Lactation and breast cancer: evidence for a negative association in premenopausal women. Am J Epidemiol 1985;121:664-674
CrossRef | Web of Science | Medline2
London SJ ,Colditz GA ,Stampfer MJ ,Willett WC ,Rosner B ,Speizer FE . Prospective study of relative weight, height, and risk of breast cancer. JAMA 1989;262:2853-2858
CrossRef | Web of Science | Medline3
Kampert JB ,Whittemore AS ,Paffenbarger RS Jr . Combined effect of childbearing, menstrual events, and body size to age-specific breast cancer risk. Am J Epidemiol 1988;128:962-979
Web of Science | Medline4
Last JM. A dictionary of epidemiology. New York: Oxford University Press, 1988:113-4.
- Citing Articles (1)
Citing Articles
1
K.B Michels, W.C Willett, D.J Hunter, G.A Colditz, B.A Rosner, J.E Manson, S.E Hankinson, F.E Speizer. (1996) Prospective assessment of breastfeeding and breast cancer incidence among 89 887 women. The Lancet 347:8999, 431-436
CrossRef
