Induced abortion and subsequent breast cancer risk: An Overview




Induced abortion and the subsequent risk of developing Breast Cancer The possibility of this association has been persistently and vigorously attacked and denied by the major medical groups in the country since l994. The ACOG, and the NCI have been particularly strong in opposing any suggestion that there is an association. In so doing, they have taken certain liberties with their interpretation of the scientific literature. AAPLOG feels that these liberties lack basic fairness and balance in reaching their “no association” conclusion. (See: Brind J; (Winter 2005) Induced Abortion as an Independent Risk Factor for Breast Cancer: A Critical Review of Recent Studies Based on Prospective Data J.Amer Physicians & Surgeons; Vol 10, #4; also see Brind J, (Sum 07)Induced Abortion and Breast Cancer Risk: A Critical Analysis of the Report of the Harvard Nurses Study II; J Amer P&S; Vol 12,#2,).

There are two pregnancy related independent risk factors for breast cancer established in the medical literature:

The first is the protective effect of an early first full term pregnancy. The landmark study establishing this protective effect [MacMahon, et al, (1970) Bull WHO 43:209-221] is widely accepted in the medical world. Obviously, aborting a first pregnancy eliminates the protective effect against breast cancer for that woman.

Using the Gail Model for Breast Cancer Risk Analysis (Thorp, et al, (2002) Long-Term Physical and Psychological Health Consequences of Induced Abortion: Review of the Evidence, Vol 58, No 1, p 75,76, and table 8,9,10), we see that an 18 year old black woman who has an abortion, then has a baby 5 years later, has double the chance of getting breast cancer than if she had delivered the first ( at age 18) pregnancy. Same scenario for a Caucasian would show a 28% increase in risk.

The second independent risk factor for breast cancer is induced abortion. As of 2004, 41 studies had been published in the worldwide medical literature (including 16 American studies) reporting data on the risk of breast cancer among women with a history of induced abortion. 29 (70%) of these studies report increased risk. Thirteen of the 16 (81%) American studies report increased risk, 8 with statistical significance (at least 95% probability that the result is not due to chance) irrespective of age at first full-term pregnancy. The relative risk increase of the 41 studies combined is 30%. (Note: this means that among aborted women there would be a 30% increase in breast cancer cases over what would normally be expected. In the current American abortion experience, this would result in approximately 5,000 additional cases of breast cancer per year in the U.S. (There are about 190,000 new cases of breast cancer diagnosed in the US each year.)

THE DALING STUDY (Janet R. Daling, Kathleen E. Malone, Lynda F. Voigt, Emily White, Noel S. Weiss, (1994) JNCI 86:1584-92l ).

The Daling study was specifically funded by the United States National Cancer Institute to investigate the abortion/breast cancer link 845 women with breast cancer were compared with 961 controls. Janet Daling’s group found an overall 50% breast cancer risk increase by age 45 for women who have had an induced abortion. A12% lifetime chance of developing breast cancer becomes an 18% lifetime chance. Among women with a family history of breast cancer (mother, grandmother, sister, or aunt), the increase in risk was 80%. If the woman had her abortion before she was18, the increase in risk was more than 100% (doubled!). If the woman had both risk factors (family history, and abortion before 18), the risk was incalculably high, i.e., there were 12 such women out of 1800 in the study, and ALL TWELVE DEVELOPED BREAST CANCER BY AGE 45. This subgroup is too small to be “statistically significant,” but surely it is “SIGNIFICANT” if you are an abortion-minded 17 year-old pregnant teenager with a family history of breast cancer!

THE HOWE STUDY (Howe et al(1989) Early Abortion and Breast Cancer Risk Among Women Under Age 40; Int J Epidemiol 18:300-4.)

This is the only study yet published on American women which relied solely on data from medical records entered at the time of the abortion (a prospective data base immune from potentially inaccurate interview material). Study includes 1451 cases, 1451 controls.

New York State kept records of fetal death (abortions) for over 20 years. This medical records based fetal death (abortion) data was collated with medical records based breast cancer data (both BrCa and fetal death were a “reportable disease,”) and matched with a control group based on driver’s license age and zip codes. The study reported a statistically significant 90% increase in breast cancer risk by age 40 with a history of induced abortion

The study was obviously immune from charges of interview bias: no interviews. Had excellent methodology: totally records based, and there was mandated reporting of both the fetal death (abortion) and of breast cancer.. The results showed a 90% increase in expected breast cancer by age 40!! Almost double the expected risk by age 40!! And how do the medical voices of denial handle this study? They don’t. They simply ignore it.


All interview studies are subject to potential inaccuracy due to “recall bias” (also called response bias, or reporting bias). But, those who deny the “ABC link” twist this concept of “recall bias” into “selective recall bias.” They theorize (without proof) that women with breast cancer will selectively admit to their abortion history, and healthy “control group” women will more likely selectively deny their abortion history. The result would be: It “looks like” abortions are associated with more breast cancer, “no abortions” are associated with less breast cancer. On this constructed theory, this undemonstrated presumption, the entire body of interview based world literature has been relegated to the “academic junk yard”.

However: three studies (Wantanabe (1968), Nippon Rishno, 26: 1853-9 from Japan, in Japanese; Lipworth (1995) Int J Cancer, 61:181-4, from Greece; Daling, (1994) JNCI 86:1584-92, USA) all had, within their study design, a method to show selective recall bias, if it existed. None showed this bias. Tang (Mei-Tzu C. Tang, Noel S. Weiss, Janet R. Daling, and Kathleen E. Malone; Case-Control Differences in the Reliability of Reporting a History of Induced Abortion; A J Epidemiology, vol 151, #12, June l5, 2000) conducted a study to specifically demonstrate recall bias. The cancer patients and the healthy controls BOTH exhibited recall bias: 14% of cases and 14.9% of controls conveniently neglected to report an abortion that they had undergone. If both the cancer patients and the healthy controls have a similar “recall bias,” this strongly suggests the world literature (29 of 41 studies by 2004) reporting an average 30% increase in breast cancer in patients who had undergone abortions is very likely accurate.


POSITIONS. (see below)

MELBYE STUDY (Melbye M, Wohlfahrt J, Olsen JH, Frisch M, Westergaard T, Helweg-Larsen K and Andersen PK (1997) Induced abortion and the risk of breast cancer. N Engl J Med336: 81–85)


Strong points:

1.Huge Data Base (all the women born/living in Denmark since 1935—about l. 5 million. But 1. 2 million have neither breast cancer nor abortion exposure).

2.Based on computerized records of abortion and breast cancer, so there is no interview problem. This should be a very reliable study.

3.The study was highly acclaimed by the NEJM, NCI, and of course by the author himself (in the Wall Street Journal: “I think this settles it. There is no overall increased risk of breast cancer” following induced abortion.

Problem points with the Melbye study:

1.Of the 10,000 women who had breast cancer, most are too old to have their abortion history on registry record, (The medical computer registry tracking abortions began in Denmark in1973.) So according to the Danish public records from 1940 to 1973 (this is NOT the Danish computerized registry, which began in l973) about 60,000 women who had abortions before 1973 were misclassified as having no abortion. And the older women (naturally) have a higher incidence of breast cancer. So this misclassified (“no record of abortion”) group is associated with “more breast cancer”, resulting in error: underestimation of the relationship between abortion and breast cancer.

2. 25% of the women in the study were under age 25 at the close of the study. Their abortions are on the government registry. But they were too young to have developed breast cancer. So “higher abortion incidence” is associated with less breast cancer, resulting in error: (again,) the underestimation of the relationship between abortion and breast cancer.

3. Now, combine these errors, as Melybe did, and the result is “no association” between abortion and future breast cancer risk.

4. The group of women who had abortions after 18 weeks showed a breast cancer rate 89% higher than controls, but this fact was diluted out when all the age groups were lumped together for the final conclusion.


The BERAL Study (Beral V; (2004) Lancet “A Collaborative Reanalysis of data from 53 studies, including 83,000 women from 16 countries.” , 363:1007-16)

This is a comprehensive meta-analysis of world literature. Proclaimed as a “full analysis of the current data,” this study found no evidence of an ABC link.

But consider how the Beral study was constructed:

By 2004 there were only 41 published world wide studies. How did she get 53 studies? By finding 12 more, one would expect. WRONG! Acually, Beral took the 41 published studies, and excluded 15 studies for reasons she felt valid (these previously published studies together showed an 80% increase in breast cancer risk after abortions!). Beral then added 27 previously unpublished studies of her own choosing, and came to the conclusion that there is no association.

Is this balanced scientific investigation?? If you are on the “choice” side of the argument, things like this get published in a prominent journal, and become fact!

How might induced abortion influence the development of breast cancer?

We don’t know for certain. But, we do know that:

1.Prior to puberty, a woman’s breast contains immature lobules, called type 1 lobules.

2.After puberty, with increasing estrogen, these lobules begin to increase in number and in maturity, and are called type 2 lobules.

3.Pregnancy produces a huge increase in estrogen levels— about 20 times nonpregnant levels. This causes an immense increase in the number of type 1 and 2 (relatively immature, in accelerated growth phase) lobules. More vulnerable lobules make more places where cancer can start.

4.In the 3rd trimester, and with lactation, the lobules complete their maturation into type 3 and 4 lobules. Studies by Russo and Russo (Jose Russo, Yun Fu Hu, Ismael D.C.G. Silva, and Irma H. Russo; Cancer Risk Related to Mammary Gland Structure and Development; Microsc. Res. Tech. 52:204 –223, 2001) demonstrate that type 3 and 4 mature lobular tissues are more resistant to cancer influences/genetic mutations than are the less mature, type 1 & 2 lobules. . The post abortive woman is left with a huge increase in the more vulnerable type 1 & 2 lobules. Thus, the process of lobular maturation in a full term pregnancy could account for “the protective effect” that is observed.

5. Abortion abruptly interrupts this process before the 3rd trimester maturation can happen by causing an immediate and marked drop in the estrogen levels. Russo notes that this leaves the type 1 and 2 lobules, now greatly increased in number, in non-mature (only partially differentiated) growth phase. He postulates that this could well make them more susceptible to malignant change with exposure to carcinogens at a future time. This could be a major factor in the increased risk between abortion and subsequent breast cancer that many studies show.

A “miscarriage” generally does not result in either high estrogen levels, nor in abrupt estrogen decrease, as in most cases the pregnancy is not progressing and not producing the expected high levels of estrogen. Most studies indicate no increase in breast cancer after miscarriage.

Induced Abortion and Subsequent Breast Cancer Risk