by: Paul Ranalli, MD

At the dawn of the 21 st century, breast cancer has emerged as a major threat to women in the Western world. In posing a lifetime risk of greater than one in 10, breast cancer now strikes over 170,000 American women and over 5,000 Canadian women every year (Ries, 1999; Cancer Centre Update, 1997). Intense interest in the search for possible genetic, dietary, and environmental risk factors is shared by the medical research community and the public at large.

Possible Abortion and Breast Cancer Link

Pathophysiology

The observed association between induced abortion and an increased risk of the later development of breast cancer is congruent with our understanding of the hormonal effects of pregnancy on a woman's breast tissue. Prior to her first pregnancy, a woman's breast is composed largely of connective tissue linking ducts which contain few milk-producing cells. Upon conception, a surge of oestradiol reaches 20-fold in the first trimester, triggering an explosive growth of breast tissue, a period when breast cells are most likely to be affected by carcinogens. When a woman completes her first full pregnancy, further hormonal changes propel these newly produced breast cells through a state of differentiation, a natural maturing process which greatly reduces the risk of future breast cancer (Kelsey, 1979, 1981; Ewertz and Duffy, 1988).

A n abrupt, premature termination of a first pregnancy by abortion arrests this process before the cancer-reducing evolution of hormone release later in pregnancy can occur, leaving a large population of dangerously-stimulated breast tissue cells in place, enormously raising future cancer risk. On the other hand, ". . . an early first full-term pregnancy would provide the greatest protection against breast cancer by drastically reducing, early on, the presence of undifferentiated and hence vulnerable breast cells, thereby decreasing the risk of subsequent transformation." (Krieger, 1989)

Animal studies support this model. Russo and Russo (1980) exposed two groups of rats to a chemical carcinogen. One group, who mated and carried a first pregnancy to term, developed mammary tumours at a rate of 6%. The other group, who mated, became pregnant, then were aborted (via hysterectomy), developed mammary tumours at an incidence of 78%; virgin rats also developed tumours at a high rate, but not as high as those aborted.

Evidence of Risk in Humans

In 1994, Dr. Janet Daling, a research epidemiologist at the Fred Hutchison Cancer Institute in Seattle, published a study in the Journal of the National Cancer Institute (Daling, 1994) revealing that women who underwent an induced abortion had a 50% greater chance of developing breast cancer than matched control women who had not previously aborted. Importantly, Daling separated out women who had suffered a spontaneous abortion (miscarriage), and found they had no increased risk of breast cancer.

Her findings were not unique. In fact, of 14 U.S. studies to date, looking specifically at the risk of breast cancer in women with a history of induced abortion, 13 of these studies have shown an increased risk. The first of these, a study by Pike and colleagues (1981), initially funded by the U.S. National Cancer Institute (NCI) and published in the British Journal of Cancer , uncovered a 137% increased risk of breast cancer. They concluded that "a first-trimester abortion . . . before first full-term pregnancy appears to cause a substantial increase in risk of subsequent breast cancer. Our finding makes biological sense if one considers breast tissue as merely proliferating in early pregnancy; the protective effect of a first full-term pregnancy is then brought about by a combination of cell differentiation and possibly permanently altered hormone levels."

These studies of American women reinforce earlier and subsequent international investigations that now total 33 epidemiological studies worldwide, of which 27 show a higher risk of breast cancer in women who have chosen abortion. The original report of Segi in 1957 (Segi, 1957) based on Japanese women diagnosed with breast cancer between 1948-52 found a 163% increased risk. A later Japanese investigation (Nishiyama, 1982) of women with breast cancer in Tokushima prefecture found a nearly identical 152% increased risk. Along with two other positive Japanese studies, women who have undergone abortion were found to have an increased risk of breast cancer in Russia (71% increase) (Dvoirin, 1978), France (32% increase) (Le, 1984), Greece (51% increase) (Lipworth, 1995), and the Netherlands (90% increase) (Rookus, 1995).

Exposure

Some in the scientific community have maintained that even a 30% increased risk of breast cancer from abortion is quite small compared to, for example, the tenfold increase in lung cancer with smoking. On the other hand, the typical smoker with lung cancer has acquired his risk by smoking a pack of cigarettes a day for anywhere from 10 to 40 years, which represents 73,000 to 292,000 cigarettes. And even after thousands of cigarettes, the risk can be partly reversed if a person quits early enough. But a measurable increased risk of breast cancer comes after just one "exposure" to abortion, and abortion is not reversible.

The overall exposure of women to abortion is enormous. Of 1,500,000 abortions in the U.S. each year, 800,000 are performed on women with their first pregnancy. At a baseline rate of development of breast cancer of 10%, a 30% increased risk above this could result in 24,000 extra cases of breast cancer per year as these women age through the next few decades. In Eastern Europe, the phenomenon may underlie the recent alarming increase in the incidence of breast cancer in younger women. While the typical breast cancer patient in Western Europe presents in her 40s and 50s, a marked increase in breast cancer onset in the late 20s and early 30s has been observed in Lithuanian women (Rich, 1994), many of whom may have had five or six abortions by their mid-20s.

Response in North America

Despite intense interest in the public and medical research community to seek possible genetic, dietary, and environmental risk factors for breast cancer, the findings of both North American and international studies linking abortion with breast cancer have generally been ignored by North American cancer research authorities. The official website of the U.S. National Cancer Institute (cancernet.nci.nih.gov) minimizes the abortion-breast cancer link, as do other national cancer institutes, such as in Canada (cancer.ca), and among public health information groups (healthlinkusa.com). Until recently (Armstrong 2000), authoritative medical reviews of breast cancer risks have not even mentioned induced abortion.

Where a connection between abortion and an increased risk of breast cancer has been discovered, researchers have often found their findings either minimized or questioned by the medical and research establishments. Before the 1994 publication in the United States of the Daling research, the Journal of the National Cancer Institute (JNCI), stated about earlier studies: ". . . recently, foes of abortion and some scientists have been pointing to a few studies that suggest that an aborted pregnancy increases the risk for the disease." Since Daling's and other research scientists' publication of data linking abortion and breast cancer, the controversy over their findings has been discussed in the mass media, including a newspaper article in the Wall Street Journal entitled "The Politics of Breast Cancer"(1997). In this article, John McGinnis noted: "Recently . . . several respected, supposedly impartial scientific researchers have been brushing aside . . . evidence of a link between abortion and breast cancer, thus allowing the politics of abortion to discourage at least one area of breast cancer research."

A clarifying meta-analysis

One of the researchers whose work has been most questioned by the cancer establishment is Dr. Joel Brind, a professor of endocrinology at Baruch College, City University of New York. At the same time Daling was publishing her results, Dr. Brind and his team were sifting through decades of published data on the epidemiology of breast cancer. Using the technique of meta-analysis to look at data from previous studies, Brind found a significant connection between induced abortion and the later development of breast cancer.

Brind and biostatistician Vern Chinchilli pooled together patients and control subjects from 28 original published reports, establishing stringent and conservative criteria to select data in which exposure to induced abortion could be separated clearly from spontaneous abortion. The overall odds ratio, for any abortion exposure, of the risk of breast cancer was found to be 1.3, a 30% increased risk (where 1.0 represents no risk). The study is so statistically powerful that the 95% confidence interval was a tight 1.2 to 1.4 (20% to 40% increased risk), meaning there is less than a one in 20 likelihood that the increased risk of breast cancer could be anything less than 20%. Statistically, Brind's study is virtually unassailable, and its publication ignited a storm of controversy.

Effect of Delayed Childbirth

A crucial feature of Brind's study was his careful separation of the independent effects of abortion on a woman's breast cancer risk from the previously-known risk of delaying her first full-term completed pregnancy. Some critics in the medical community assert that this delayed first-birth effect is the only explanation for a claimed increased risk of breast cancer. Brind's study conclusively demonstrates that abortion is an independent risk factor in its own right. This conclusion was achieved in two ways: (1) By including studies in which the control group (women without breast cancer) included nulligravid women (women who had never been pregnant); and (2) In studies where some women had given birth, at various ages, a calculation was performed to subtract out the effect of age at first live birth. The result? The independent risk of induced abortion was still significant.

Brind and his colleagues argued that the two effects - delayed childbirth and abortion -- are additive. If it is assumed, conservatively, that an average woman's lifetime risk of breast cancer is 10% (one in 10), it is known that an early full-term pregnancy reduces this risk, from 10% to about 7%. If a young pregnant woman opts instead for an abortion, she relinquishes the benefit of an early completed pregnancy and, in addition, adds the independent 30% increased risk from the abortion, raising her risk from 10% to 13%. Thus the decision to abort her first pregnancy will nearly double her lifetime risk of breast cancer, from 7% to 13%. A second abortion will add further risk, both from the abortion itself and by further delaying the protective effects of a first completed pregnancy.

Despite the statistical power of Brind's study (or, perhaps, because of it), many in the medical and scientific community were quick to rebut his findings. The New England Journal of Medicine published a remarkably flawed Danish study (Melbye, 1997) which explained away a 44% increased risk of breast cancer in women with a history of abortion as being based on an otherwise unexplained global increase in breast cancer incidence. The JNCI offered a generic criticism of the technique of meta-analysis, stating (JNCI 1997) that "biased studies entered into a meta-analysis produced biased results."

With time, however, the quality of Brind's study has gradually begun to win grudging acceptance among important sectors of the world medical community. In April 2000, Britain's Royal College of Obstetricians and Gynecologists (RCOG) published Evidence-based Guideline No. 7: The Care of Women Requesting Induced Abortion , which said of two of the most thorough reviews of the abortion-breast cancer literature, one being Brind's meta-analysis: "These two meta-analyses were independently assessed for the RCOG Group. The assessor concluded that both were carefully conducted reviews and that the Brind paper had no major methodological shortcomings and could not be disregarded."

Even more significant was the inclusion, for the first time, of abortion as a risk factor for breast cancer in a February 2000 review of the subject by Katrina Armstrong and colleagues in the New England Journal of Medicine (Armstrong, 2000). Although abortion was downplayed as one of four "risk factors . . . less consistently associated with breast cancer", its inclusion in such a short list represents a significant acknowledgement. A 1992 review in the same journal (Harris, 1992) did not mention abortion, despite 45 years of evidence at that point.

Recall Bias

One way of explaining a clear emerging worldwide trend linking abortion with an increased risk of breast cancer, is the concept of recall bias, proposed by Harris and colleagues (1989). These authors postulated that "a woman with cancer is perhaps more likely to remember and report a previous abortion than a healthy control". If this was true, a falsely elevated apparent risk in breast cancer patients might result. The only support for this notion rests with a set of Swedish data (Harris, 1989; Meirik, 1986) which shows that, rather than non-cancer patients under-reporting abortions, several women with cancer seemed to over -report abortions, that is, apparently they declared abortions they never had (based on discordance between a computerized registry and interview data). Few workers in the field accept this concept, which raises questions about the quality of data reporting in that study. Indeed, the Swedish authors of this study eventually retracted their claim (Meirik, 1998). In any event, there are now four studies (Watanabe, 1968; How, 1989; Daling, 1994; Lipworth, 1995) whose design has conclusively ruled out any evidence of recall bias.

The importance of identifying precise studies

Over two dozen other studies of the association of abortion and breast cancer since 1960 are betrayed by various confounding factors which prevent an examination of uncontaminated data related to induced abortion. The most common error is the failure to separate data from women who have suffered miscarriages ("spontaneous abortion") from those who underwent induced surgical abortion. Spontaneous abortion has long been recognized to offer no increased risk of subsequent breast cancer, and there are clear biological reasons for this. It appears that miscarriage pregnancies are doomed from the earliest days after conception by a failure to develop the expected estrogen hormonal surge, thus these women are never exposed to the powerfully high estrogen levels of a healthy pregnancy that is abruptly terminated by abortion. The low maternal estradiol surge in spontaneous abortion was first observed by Kunz and Keller in 1976 (Kunz and Keller, 1976), and has recently been confirmed by Stewart and colleagues (1993). This biological difference between spontaneous and induced abortion underlies the flaw inherent in epidemiological studies that pool data from both groups, and points out the value of a meta-analysis such as the one carried out by Brind (Brind, 1996), which isolates and studies data from induced abortion alone.

A study that mixes spontaneous abortion cases with induced abortion is imprecise, and results in a falsely low apparent risk of subsequent breast cancer. For example, in a 1996 study by Newcombe and colleagues, which garnered prominent attention after its publication in the Journal of the American Medical Association (Newcombe, 1996), the actual increased breast cancer risk of women exposed to induced abortion was 23%, but by mixing in women who had suffered miscarriages, the risk was watered down to 12%. It was this 12% figure that was most quoted in press reports, misleading the public into believing the study showed only a minimal risk of breast cancer from abortion. It is clear that further human studies are needed which separate miscarriages from induced abortions in order to further our understanding of this important area of women's health research.

Age at First Abortion

Only a few studies (Howe 1989; Daling 1994) have looked at the question of whether a woman who is aborted at a very young age faces a significantly higher risk of developing breast cancer. In noting that the rate of cell proliferation is likely to be highest in the youngest subjects, Daling and her colleagues have suggested that the greater risk for women younger than 18 at the time of their first abortion may be real, and should be further investigated.

Family History

Although data on this subject is limited, the observations are ominous. In the Daling study of 2,000 women, 12 women had a combination of a positive family history of breast cancer and an abortion before age 18. All 12 women went on to develop breast cancer before the age of 45. The risk in this study was, therefore, incalculably high.

Summary

There are serious grounds for believing that induced abortion creates an increased risk of breast cancer, and that this risk may be more significant for women with a positive family history. Research studies in this field have been hampered by the omission of key information, imprecise gathering of data (mixing miscarriages with induced abortions), and the politicized nature of the subject, all of which conspire to create significant barriers to a true understanding of the risk. Many North American researchers who study breast cancer are unwilling to accept induced abortion as a factor worthy of study.

For women considering abortion to have the benefit of a truly informed choice, a major shift in the medical paradigm is required. It has long been observed that the medical establishment is slow to respond to emerging data. The smoking-lung cancer link, and the relationship between diet and health, are but two examples of concepts which have taken years, even decades, to become accepted, indeed form part of conventional medical wisdom. Fortunately, the internet age is likely to increase the availability and transparency of medical information to the public, ultimately liberating primary data linking abortion and breast cancer from the government and medical hierarchies who may be subject to political and ideological influence. Women will thus be allowed to directly inform themselves of the risks to their health posed by induced abortion.

World epidemiological studies on the association of breast cancer with induced abortion.

Risk of abortion presented in terms of Odds Ratios (OR), ie. OR=1.3 represents a 30% increased risk of breast cancer; OR=0.9 represents a 10% reduced risk; OR=1.0 represents no particular risk. Variability of the data is represented by the 95 per cent Confidence Intervals (95% CI), ie. The true result has a 95% cnace of falling within the described range. If the range is beyond 1.0, the data in that particular study are statistically significant.

For further information on this topic, see Women have a right to know...