In referring patients for induced abortion, the role of a GP in providing patients with unbiased information on risks may be particularly critical. In contrast, for example, the cardiovascular risks attendant upon eating fatty foods or smoking cigarettes are widely disseminated by the lay media and public health agencies alike. And if one’s coronary arteries should succumb to such stresses, the cardiologist and/or the hospital will take care to inform the patient of all the attendant risks of angioplasty, bypass surgery, or whatever medicine and/or surgical procedure may be indicated. But abortion is so contentious an issue, and maintaining its reputation as "safe" such a popular political imperative, that unbiased information on attendant risks is rarely forthcoming, even from the usually reliable sources.

A most relevant case in point is the effect of abortion on a woman’s subsequent risk of breast cancer. The Anti-Cancer Council of Victoria states, in its Canstat booklet: "Women who are nulliparous or who postpone their first full term pregnancy to later in their reproductive life have an increased risk of breast cancer." While the relevance of this statement to an abortion decision is obvious to a physician—that a young nullipara who chooses abortion over childbirth will thereby increase her subsequent risk of breast cancer—its significance to a patient seeking a procedure not even named therein is hardly so. Moreover, even the booklet itself would likely be found only in the offices of oncologists—and GP’s—but not in the offices of abortion providers. That is tragic, considering that the risk-enhancing effect of delaying first full term pregnancy has been widely acknowledged since the publication—in 1970—of an international, multi-center study by the World Health Organization.

Yet more tragic it is that abortion—in addition to delaying first full term pregnancy—adds independently to breast cancer risk for women in general, regardless of age or parity. I and my co-authors at the Pennsylvania State University College of Medicine published our "comprehensive review and meta-analysis" of the induced abortion-breast cancer literature in 1996, in the British Medical Association’s Journal of Epidemiology and Community Health. In this paper we reviewed all 23 then extant studies (currently there are 31) which presented specific data on induced abortion (as opposed to spontaneous abortion or unspecified abortion) and the incidence of breast cancer. We found that 17 of the 23 studies (now 25 of 31) had reported elevated risk, with an overall, statistically significant risk increase of 30%. While 30% is an epidemiologically small elevation, the incidence of breast cancer is so high—cumulative lifetime risk now exceeding 10% among Australian women—that a 30% increase in incidence, in addition to abortion’s effect in delaying first full term pregnancy, may raise the annual number of excess Australian cases into the thousands.

In light of such daunting statistics—which date back as far as 1957, and include studies on women of Asian and African as well as European ancestry—it is indeed astonishing that such a highly placed physician as Dr. Paul Niselle of the Medical Indemnity Protection Society should dismiss the idea that physicians need to warn women considering abortion about increased breast cancer risk as "a nonsense—it’s so remote" (this journal, 13 Nov., 1998). Indeed, in the one study of Australian women (originally published in 1988 in the American Journal of Epidemiology, without the abortion data, which were finally published in 1995 in the British Journal of Cancer), those who had chosen abortion were found to have a 160% increased risk of breast cancer.

Nevertheless, as the current firestorm of public debate about such breast cancer warnings in the ACT exemplifies, devotees of "safe abortion" seem bent on explaining away the connection between abortion and breast cancer, as if it were a statistical artifact arising from response bias in retrospective studies. This idea—Originally introduced by Swedish researchers in 1989—holds that breast cancer patients would be more forthcoming at interview about their abortions than would healthy women (controls), resulting in an apparently higher frequency of abortion among them, which would in turn translate into an artificially elevated relative risk. In fact, the only significant evidence in favor of this hypothesis has since been retracted by its original proponents (in the March, 1998 Journal of Epidemiology and Community Health).

Even so, one might well ask whether a mere statistical association justifies raising the spectre of cancer for patients seeking abortion, absent a reasonable biological mechanism to explain the effect. Indeed it may seem counter-intuitive that a single, simple surgical procedure such as induced abortion could have a measurable effect in the risk of cancer years, or perhaps even decades later. But the biological basis for an abortion-breast cancer link is far from absent.

Most identifiable risk factors for breast cancer are attributable to some form of excess exposure to some form of oestrogen. Thus, higher risks are seen in women with early menarche, late menopause, fewer children: more menstrual cycles with their monthly surges of ovarian oestradiol. Even such non-reproductive risk factors as chronic alcohol consumption and postmenopausal obesity are explicable in terms of chronically higher levels of endogenous oestrogens. Since oestrogens are the main growth promoters of normal as well as malignant breast tissue, the cancer connection is not a mystery. What is different about the artificial termination of a normal pregnancy is the enormity of oestrogen exposure. By the end of the first trimester, maternal levels of oestradiol are twenty-fold higher than at the time of fertilization—seven fold higher than preovulatory peak levels—and still on the way up. This results in maximal proliferation of undifferentiated breast tissue: more cells with cancer-forming potential. Only in the third trimester—if there is one—do other hormones differentiate this tissue for the task of milk production, permanently turning off their proliferation—and cancer-forming—potential. Consequently, full-term pregnancies lower breast cancer risk, while aborted pregnancies raise risk.

In contrast, most pregnancies which end in spontaneous abortion are characterized by subnormal secretion of oestradiol by the ovaries (in parallel to subnormal secretion of progesterone, which defect causes the miscarriage), and are not found to increase breast cancer risk, according to most epidemiological studies. Moreover, the risk-enhancing effect of artificial pregnancy termination has been demonstrated in laboratory experiments in rats, as far back as 1980.

Where does all this leave the general practitioner? The GP is in the critical—albeit uncomfortable—position of being his or her patients’ only likely source of critical information about the long-term cancer risk of a procedure for which the patient is referred, since abortion providers have amply demonstrated their unwillingness to do so.

But there is more. The GP may also in the critical position of being his or her patients’ only likely source of critical information about the patient herself, information which bears heavily on the magnitude of the likely effect of abortion on her future risk of breast cancer. Only a handful of epidemiological studies have examined the interaction of family history of breast cancer—usually the strongest risk factor observed—and induced abortion. And the effect is apparently synergistic. Thus, a 1994 French study observed, among women with a positive family history, a 610% increase in women with two or more induced abortions; a 1994 American study found an 80% risk increase among any women with a positive family history, and a 1997 American study found a 220% increase in nulliparous women with a positive family history. Most ominous was the finding, in the same 1994 American study, of an incalculably high relative risk among women with a positive family history who had had an abortion before age 18. (All 12 study subjects in this category were in the patient group—none were in the cancer-free control group.)

How often occurs the situation, in a general practice, where a teenage girl seeks an abortion referral, perhaps unaware herself of what her GP knows, that her grandmother died many years ago of breast cancer? It is indisputable that her risk of future breast cancer is already elevated by her family history, and that completion of her pregnancy will lower it. Knowing the epidemiological evidence which indicates that aborting her pregnancy will further substantially raise her breast cancer risk, perhaps to a certainty, what physician would fail to warn her of this risk, and even to dissuade her from taking it?

Finally, in regard to potential legal liability for medical negligence, the substantial increases in breast cancer risk accruing to induced abortion stand in stark contrast to the Australian legal standard established by Rogers v. Whitaker. If a physician is obligated to warn patients of a risk as remote as one in 14,000, on what possible basis can the failure to warn of the abortion-breast cancer link be singled out as an exception to the normal legal duty and characterized as "a nonsense"?