Abortion and the Risk of Breast Cancer
(Reprinted with permission from Chisholm Health Ethics Bulletin, Winter 2003)
The rate of breast cancer among American women is about one in eight. Texas has just passed a law requiring doctors to inform women seeking an abortion that it may increase the risk of breast cancer. This article discusses the evidence of this risk and the ethical need to be objective in reporting any causal link.
Risk Factors Associated with Breast Cancer
While there are many factors associated with increasing risk of breast cancer, only a few are established beyond doubt. One is age, with the risk increasing with age, especially for women over 55 years. Another undisputed factor is a family history of breast cancer, especially in the maternal line. Research shows that the BRCA1 and BRCA2 genes are not only implicated in the onset of ovarian and colon cancers but are also associated with breast cancer. A third factor is a higher incidence of breast cancer in nulliparous women.
Women are at 150 times greater risk of invasive breast cancer then men. A statistically significant increased risk of breast cancer is associated with the early onset of menstruation prior to the age of 11 years, the late onset of menopause in women over 55 years and first fullterm childbirth after the age of 30 years. Another correlate of increased risk derived from epidemiological studies is age at first pregnancy.
'the risk of breast cancer in women with a history of induced abortion was not different from that of women without such a history
Factors that Increase Risk
Since the 1950s, there has been epidemiological evidence supporting a positive association between induced abortion and breast cancer in women. During the ensuing decades the presumption favoured a positive link, but the debate continued since the evidence did not seem to be consistent across all studies, nor did it convince the scientists and clinicians involved. In 1996 an important article reviewing multiple research studies to see if there was, indeed, an increased risk for breast cancer attributable to induced abortion was published by Joel Brind and his colleagues. It was based on a thorough meta-analysis of 28 reports of case-control studies published over the previous 40 years. The researchers believed their conclusions were valid and that bias in self-reporting was not an issue. They suggested that the surging levels of estrogen in the first trimester could significantly increase the risk of breast cancer if the pregnancy is aborted. They found that women who had an induced abortion were 1.3 times more likely to contract breast cancer than women who did not have an abortion. They concluded that their results 'support the inclusion of induced abortion among significant independent risk factors for breast cancer, regardless of timing of abortion relative to the first term pregnancy.'  This study boosted the presumption in the public perception that there was a significant link between induced abortion and breast cancer. As a result of this study and out of a sense of duty to warn women of this link between induced abortion and breast cancer, the Fact Sheet posted on web site of the US National Cancer Institute (NCI) mentioned this increased risk factor until March 2002.
Overall Link between Abortion and Breast Cancer Questioned
In 1997 Mads Melbye and his colleagues published their study of all Danish women born between 1 April 1935 to 31 March 1978 with the relevant information on the number, dates and gestational age of each induced abortion obtained by linkage from the National Registry of Induced Abortions. Likewise by linkage with the Danish Cancer Registry, all new cases of breast cancer were identified. From a total of 1.5 million women, they found there were 370,715 induced abortions among 280,965 women, of whom 10,246 were identified to have breast cancer. They found 'the risk of breast cancer in women with a history of induced abortion was not different from that of women without such a history, after potential confounding by age, parity, age at delivery of first child, and calendar period was taken into account' -- the risk odds ratio being slightly less, 1 against 1.06. A similar Swedish study on all women born in Sweden between 1 January 1973 and 31 December 1991 was published in 2003 by Gunnar Erlandsson and colleagues and it concluded in much the same vein: ' our study strengthens the evidence that neither induced not spontaneous abortions increase the risk of breast cancer.’
Quest for a Solution
An analysis was made of a case-control study of the association of induced abortion and breast cancer using prospectively recorded exposure information of women born between 1973-1991 from the Swedish Medical Birth Register. Cases were identified by linkage with the Swedish Cancer Registry and controls were randomly chosen from the birth register. The study showed overreporting of induced abortion by the cases compared to extensive underreporting of abortion by controls. Apparently women who know that they have breast cancer are more likely to reveal that they have had an induced abortion than controls who do not if they have breast cancer. Naturally self-reporting has been found to be less reliable. This is understandable, granted the likelihood of inaccuracies relative to the data of induced abortions due to the memories, emotions and social stigma involved. This same reporting bias has also been found in the case of abortion surveys. This is the main criticism of Brind's meta-analysis: if reporter bias flawed the original studies, this bias would also flaw his results. Both the Melbye and Erlandsson studies relied on data from the relevant national registers and not on self-reporting and they included all women born in the time-frame of the studies. Prospective epidemiological studies are considered by many to be more accurate and robust than retrospective research. The view that induced abortion has little influence on breast cancer is confirmed by a study of Chinese women who have induced abortions without feeling stigmatised.
Brind and Vernon Chinchilli countered by pointing out serious flaws in the above mentioned Swedish and Danish studies: their short follow-up period of younger women and misclassification errors due to the use of birth registers from 1935-1978, cancer registers only from 1968 and the abortion registry from 1973 whereas abortion had been legalised in Denmark since 1939, resulting in 60,000 elderly women who had abortions being classified as abortion negative. Brind and Chinchilli claim these faults account for the striking 'underestimation of the real induced abortion association in Denmark’.
Granted the uncertainty of the association of induced abortion with breast cancer, the National Cancer Institute convened a meeting of more than 100 world experts involved in this area to attend The Early Reproductive Events and Breast Cancer Workshop, 24 - 26 February 2003. Participants included breast cancer experts, epidemiologists, clinicians, basic scientists and breast cancer advocates. They evaluated the evidence of indications of pregnancy related to cancer, the biological changes resulting from pregnancy that may influence the risk of breast cancer and the biological mechanisms identified from animal studies. As a result of this Workshop it was recognised that the following epidemiological findings were well established:
After reviewing all the available evidence its concluding recommendation is that ‘available evidence on an association between induced abortion and breast cancer is inconclusive'
The NCI's Boards of Scientific Advisors and other Scientific Counsellors unanimously approved these findings on 3 March 2003. The Workshop admitted there were both epidemiological and clinical gaps yet to be researched and bridged. They also detailed directions for future research. The March 2003 NO Fact Sheet was accordingly updated and stated that 'having an abortion or miscarriage does not increase a women's subsequent risk of developing breast cancer.'
Brind was present at this Workshop but says no provision was made for open discussion of opposing views. He was disappointed and published a minority report which is highly critical of the procedures adopted at the Workshop. His theory is based on the fact that estrogen has a proliferative effect on breast tissue as it stimulates ductal growth. Most breast cancers are ductal carcinomas. Brind refers to the 'role of estrogen as a stimulator of cellular proliferation, as well as the known genotoxic effects of certain estrogen metabolites'. Since the levels of estrogen are very high during the first two trimesters of pregnancy, an induced abortion would deprive women of the protection derived from high levels of progesterone present at full-term pregnancy.
Some elements of Brind's theory are confirmed by the conclusion of Isha Mustafa and Kirby Bland that ‘estrogen must be an influential and essential hormone to breast growth and tumorigenesis' and that women ‘who underwent induced abortions at 12 weeks or greater gestation had an increased risk for breast cancer development' compared to those who had an abortion at less than 7 weeks. Mustafa and Bland cite studies which suggest a mechanism for reducing the risk of breast cancer: progestin, which is produced late in pregnancy, decreases estrogen-stimulated growth of breast cells and promotes differentiation which thereby lessens the risk of cancer. It seems Brind is right to say 'the great surge of the hormone estrogen could indeed be responsible for increased risk if the pregnancy is aborted. ... and that it is a matter of settled science ... that full-term pregnancies lowers a woman's long-term risk of breast cancer.' His theory is also supported by the conclusion of an unrelated study on oral contraceptives (OCs) and breast cancer: 'Our findings suggest that the increased risk of breast cancer related with [the] total duration of OC use is due mostly to [the] estrogen component.'
The U.K. Royal College of Obstetrics and Gynaecology in March 2000 published evidence based guidelines on 'Care of women requesting induced abortion', and is now preparing an opinion paper on abortion and breast cancer. After reviewing all the available evidence its concluding recommendation is that 'available evidence on an association between induced abortion and breast cancer is inconclusive'. The specific point was made that ' Brind's paper had no methodological shortcomings and could not be disregarded.'
The risk factors for breast cancer are many and varied; it is difficult to unravel all its intertwined causes. It seems more prudent to accept the opinion of the U.K. RCOG that the link between abortion and breast cancer is inconclusive. In the light of all the evidence, it would be true to say that women who have an induced abortion during their first and only pregnancy would certainly be at an increased risk of breast cancer. It is also true to say that if women who have had an abortion were to have one or more subsequent full-term pregnancies, they would benefit from a reduced risk of breast cancer. In short, while full-term pregnancies do decrease the risk of breast cancer, an induced abortion may well deprive a woman of some protection against breast cancer, but not significantly increase its risk.
The fact that some in the community, myself included, morally disapprove abortion, while others approve it, does not alter the ethical imperative to provide accurate information on the increased risk of breast cancer, if any, due to induced abortion. If the evidence suggests there are reasonable grounds for an increased risk, this should be communicated to women seeking an abortion. It is, then, necessary to address and resolve the remaining differences found in studies on this matter. More flawless research should provide the answers. Women have a right to know the truth, and fears that
induced abortion may not appear safe should not be allowed to obscure the facts.
Norman Ford SDB