The single most avoidable
for breast cancer is...
The ESTROGEN Connection:
The type of female sex hormone called estrogen, is the most potent stimulator of breast
cell growth. In fact, the actions of most known risk factors for breast cancer are
attributable to some form of estrogen overexposure.
In a normal pregnancy, the mother's ovaries begin producing extra estrogen within a few
days after conception 38.
The level of estrogen in her blood rises by 2,000% by the end of the first trimester-to a
level more than six times higher than it ever gets in the non-pregnant state 39,40.
It is the undifferentiated cells in the breasts which estrogen stimulates to proliferate,
so that there will be enough milk-producing tissue to feed the baby after birth. Only the
undifferentiated cells are vulnerable to carcinogens, and can ultimately grow into cancer
Importantly, during the last 8 weeks of pregnancy, other hormones differentiate these
cells into milk-producing cells. In the process the growth potential-and cancer-forming
potential-of these cells is turned off. That is why a full-term pregnancy lowers the risk
of breast cancer later in life 41.
Therefore, if a woman who has gone through some weeks or months of a normal pregnancy
chooses abortion, she is left with more of these cancer-vulnerable cells in her breasts
than were there before she got pregnant, which raises her risk of breast cancer later in
In contrast, most pregnancies which abort spontaneously do not generate normal quantities
of estrogen 39,40. Thus most miscarriages
(at least 1st trimester miscarriages) do not raise breast cancer risk 36.
a. Mature breast of a never-pregnant woman
Summary and meta-analysis of epidemiological
evidence of the abortion-breast cancer link:
b. Breast at end of full-term pregnancy
Each study above is listed by first author's name, year of publication,
and nationality of women studied. On the right-hand side of the figure, the horizontal
line with a central circle, given for each study, represents (on a log scale) the 95%
confidence interval (CI) for the effect of induced abortion on the entire population
studied, with the central circle representing the "point estimate" of
"relative risk" (RR). This RR value represents how many times more likely to
develop breast cancer, in that particular study, is a woman who had at least one induced
abortion, relative to a woman who has not had an induced abortion. For example, in the
1984 French study of Le et al shown above, the point estimate of RR is 1.5, with a 95% CI
that spans from 1.0 to 2.2. In other words, the study found that women who had at least
one abortion were, on average, 50% more likely to develop breast cancer, and that one can
be 95% certain that the increased risk is between 0% and 120%.
Point estimates to the right of the vertical line of unity (RR=1)
indicate increased risk; while those to the left indicate decreased risk. If the 95% CI
does not cross the line of unity, the results are said to be statistically significant.
Narrower 95% CI's denote greater certainty about the RR value, reflecting larger studies
with greater statistical power. Thus, the figure illustrates the fact that of the 33
published worldwide studies, 27 show increased risk, 17 of which are statistically
significant. The pooled average from all the studies combined, calculated by two different
methods, is shown at the bottom. It clearly indicates a significant risk increase
averaging 30 to 40%.
The 1981 study of Pike et al. is limited to women with any abortions
before first full-term pregnancy, the 1988 study of Ewertz and Duffy and the 1996 study of
Wu et al. are limited to women with no children, and the 1957 study of Segi et al. is
limited to women with children.
References (To view or print this list of references, click here to open a PDF file.
1. Pike et al. (1981) Br J Cancer 43:72-6
2. Brinton et al. (1983) Br J Cancer 47: 757-62
3. Rosenberg et al. (1988) Am J Epidemiol 127-981-9
4. Howe et al. (1988) Int J Epidemiol 18:300-4
5. Laing et al. (1993) J Natl Med Assoc 85:931-9
6. Laing et al. (1994) Genet Epidemiol 11:A300
7. White et al. (1994) J Natl Cancer Inst 86:505-14; Daling
et al. (1994) J Natl Cancer Inst 86:1584-92
8. Newcomb et al. (1996) JAMA 275: 283-7
9. Daling et al. (1996) Am J Epidemiol 144:373-80
10.Wu et al. (1996) Br J Cancer 73:680-6
11.Palmer et al. (1997) Cancer Causes Control 8:841-9
12.Marcus et al. (1999) Am J Pub Health 89:1244-7
13.Lazovich et al. (2000) Epidemiol 11:76-80
14.Moseson et al. (1993) Int J Epidemiol 22:1000-9
15.Segi et al. (1957) GANN 48 (Suppl.):1-63
16.Watanabe & Hirayama (1968) Nippon Rinsho 26:1853-9
17.Dvoirin & Medvedev (1978) Meth Breast Cancer Epidemiol
Res, Tallin 1978. USSR Acad Sci pp.53-63 (in Russian)
18.Nishiyama (1982) Shikoku Ichi 38:333-43 (in Japanese)
19-22. Le et al. (1984); Luporsi (1988); Rohan (1988);
Andrieu et al. (1994); in Andrieu et al. (1995) Br J Cancer
23.Hirohata et al. (1985) Natl Cancer Inst Monogr 69:187-90
24.Ewertz & Duffy (1988) Br J Cancer 68:99-104
25.Lipworth et al. (1995) Int J Cancer 61:181-4
26.Rookus & van Leeuwen J Natl Cancer Inst 88:1759-64
27.Bu et al. (1995) Am J Epidemiol 141:S85
28.Talamini et al. (1996) Eur J Cancer 32A:303-10
29.Burany (1979) Jugosl Ginekol Opstet 19:237-47 (Serbo-Croat)
30.Adami et al. (1990) Br J Cancer 62:122-6
31.La Vecchia et al. (1993) Int J Cancer 53:215-9
32.Zaridze et al. (1988) "unpublished" in Ref.
33.Melbye et al. (1997) N Engl J Med 336:81-5
34.Rosenberg (1999) NE FL Women's Health v. State of FL,
FL Circuit Ct, 2nd circ., videotape deposition of 11/18/99,
35.Clark & Chua (1989) Clin Oncol 1:11-18
36.Brind et al. (1996) J Epidemiol Community Health 50:
37.Evidence-based Guideline #7 (2000) RCOG Press, pp.29-30
38.Stewart et al. (1993) J Clin Endocrinol Metab 76:1470-6
39.Witt et al. (1990) Fertil Steril 53:1029-36
40.Kunz & Keller (1976) Br J Ob Gyn 83: 640-4
41.MacMahon et al. (1970) Bull Wld Health Org 43:209-21
BREAST CANCER PREVENTION INSTITUTE
Call toll-free: 1-86-NO CANCER (1-866-622-6237)