Heredity, gender and the discourse of ovarian cancer

Abstract

Healthy women have elected to undergo preventive, or “prophylactic” surgery in order to lower perceived risk for ovarian cancer for some time. The development and commercialization of cancer susceptibility tests has altered the justification for prophylactic surgery. As a result, the rationale for undergoing oophorectomies (the removal of the ovaries) has taken on greater significance in biomedical discourse as a reasonable course of action for women who have been told they have a significant risk for these cancers. Although new in many respects, cancer susceptibility genetic tests are incorporated into old frameworks for defining, understanding, and treating disease. This essay addresses what some of those discursive/material conditions are, in part by addressing fundamental assumptions that inform the ways in which genetic medicine is interpreted and crafted in response to women's cancers. BRCA research, I argue, represents the powerful intersection of heredity and gender bias in biomedicine. The result is a new, “rational” discourse of diseasing women's reproductive organs.

In the mid-1990s, a woman known only as ‘A.H.’ in a case report (Spear et al., 1999) endured a cancer-related operation lasting approximately 11 hours. During the procedure, surgeons removed A.H.'s breasts, ovaries and uterus, and plastic surgeons reconstructed breasts by using tissue from her abdomen.

A.H. did not have cancer. She did, however, test positive for a mutation of the breast cancer susceptibility gene known as BRCA2 (BRCA is short for ‘breast cancer’) and knew of three female family members with cancer. After consulting with unspecified specialists regarding her perceived risk of getting breast and ovarian cancer, A.H. underwent a risky, experimental procedure, performed only a handful of times at the medical centre in which she was a patient. According to the authors of her case report, the surgery was expected to increase her life expectancy by four to six years ‘or more’ (Spear et al., 1999).

The obvious question raised by this case report is, why would a healthy woman undergo the removal of so many body parts in order to lower her risk for cancer? Healthy women have, in fact, elected to undergo preventive or ‘prophylactic’ surgery in order to lower perceived risk for cancer. The prophylactic surgery phenomenon can be explained in part by the fact that cancer is, for many people, a frightening disease. Not only do a great number of women die from cancer each year, cultural historians have suggested that fear of cancer may exceed the real threat it poses to the general population (e.g. Patterson, 1987). Moreover, pedigree analyses of ‘cancer syndrome families’ suggest that cancer risk should be taken very seriously for some women; prophylactic surgery has no doubt extended the lives of many women.

That A.H. chose prophylactic surgery to reduce her risk for cancer is not exceptional. Even so, the development and commercialization of genetic tests of cancer susceptibility has changed the justification for prophylactic surgery, or to put it a different way, the evidentiary grounds on which it is recommended in the cancer literature have shifted. Tests for mutations in the BRCA1 and BRCA2 genes have captured the attention of geneticists and cancer researchers since mutations in these genes are strongly linked to an increased risk for breast and ovarian cancer and offer insight into the mechanisms by which risk is transmitted from one family member to another. Genetic tests provide a type of evidence for inherited susceptibility that is qualitatively different from risk assessment gleaned from family pedigree analysis, the latter providing risk information that is relevant only for women who have in fact inherited a defective mutation in a cancer susceptibility gene such as BRCA—something that only a genetic test can confirm. BRCA tests, then, can detect whether a woman has inherited a mutation and clinical trials provide data regarding the risk associated with a particular mutation (e.g. the 185delAG of BRCA1) or type of mutation (e.g. protein-truncating). Moreover, the meaningfulness of BRCA tests for the prevention of cancer increased significantly with the publication in 1999 and 2002 of studies demonstrating the efficacy of mastectomy and oophorectomy (removal of the ovaries) for reducing risk for both breast cancer and ovarian cancer (Hartmann et al., 1999; Kauff et al., 2002; Rebbeck et al., 2002). For cancer researchers and practitioners, these studies provide clinical evidence that surgery can be reasonably expected to manage risk that only BRCA tests can quantify with any certainty.

For feminists, the important question is not merely how do we detect and manage risk, but rather, what do we mean by ‘risk’ and who inhabits its categories? Indeed, the development of genetic screens and the legitimization of prophylactic surgery have accompanied transformations in the ways in which cancer researchers conceptualize risk. First, the subject of risk has shifted from the cancer syndrome family to any woman with a BRCA mutation, regardless of family history. Second, what counts as—for lack of a better word—‘actionable’ risk has shifted. In pre-BRCA risk analysis, only women in cancer syndrome families were reasonably expected to consider prophylactic surgery. In the wake of commercialized BRCA tests and clinically relevant surgical procedures, a consensus is slowing emerging regarding prophylactic surgery as a rational, commonsense option for all women with inherited genetic mutations. And as researchers shift their gaze from family-specific to population-specific mutations, the number of women for whom these recommendations are relevant will only increase.

This article offers an analysis of these shifts in discourse, the latter defined as the totality of terms, tropes, ideas and arguments which codify the belief among cancer researchers that BRCA tests reveal quantifiable, modifiable risk. Here we see rhetoric as an obvious alternative to a correspondence theory of language—rhetorical language mediates the scientist's access to the object world, bringing to bear habits of mind grounded in socially and culturally significant belief systems. Science studies scholars hailing from gender studies, cultural studies and continental philosophy of science have noted that scientific discourse is a privileged one, claiming to represent the ‘real’ while always already articulating, legitimizing and, indeed, naturalizing, social structures and the subjectivities they create and enable. Nowhere is this more apparent than in the biological sciences, in which the object of study is the human body, understood along the axes of gender, race and sexuality.

Feminists understand that the body is not ‘mere’ discourse, that the corporeality of the body resists discourse as much as it is affected by it. The performance of oophorectomy and mastectomy is an event wherein representations matter—risk, as knowledge, translates into the literal transformation of a woman's body. BRCA mutation carriers both inhabit the risk categories that genomics enables and withstand surgical—and in some cases pharmacological—interventions, and it is in these ways that women arguably embody heredity. Feminist science scholars such as anthropologist Emily Martin have shown that women literally embody medical discourse—metaphors of waste and decay affect the ways in which women talk about, and in turn experience, events like menstruation and menopause. The discourse of risk that marks the literature on inherited ovarian cancer likewise impacts on the experience of risk. Yet it also informs decisions regarding its management—in the case of oophorectomy, the decision to remove healthy ovaries. Since oophorectomy in healthy women predates the commercialization of BRCA mutations screens, genomics cannot explain entirely why physicians and cancer specialists recommend this procedure for women at risk. Rather, one must situate the development of reproductive cancer genomics within a broader socio-cultural context in which researchers bring to bear habits of mind about women, reproduction and motherhood.

In this article, I focus on the discourse of inherited ovarian cancer and the role of oophorectomy as prevention for ovarian cancer risk. In addition to the widely discussed link between BRCA mutations and breast cancer risk in public discourse, cancer geneticists have linked the BRCA genes to an increase in ovarian cancer risk anywhere from 10% to 60% over the course of a woman's lifetime. Fewer cases of ovarian cancer are diagnosed each year compared to breast cancer, but the vast majority of those new cases are diagnosed during the advanced stages of the disease during which options for successful treatment diminish. Thus there is an urgency with which researchers talk about the need for women to consider seriously having their ovaries removed by their mid-30s or ‘when childbearing is complete’. As many women are not done with childbearing by this time, and removal of ovaries when a woman is pre-menopausal presents health risks of its own, developments in inherited ovarian cancer research are of immeasurable interest to feminist science studies scholars and the women whose livelihoods are affected by developments in cancer genomics.

In what follows, I offer a description and explanation of the discursive shifts that make possible the diseasing of women's bodies in the absence of actual cancer, examining a representative sample of research reports and editorials published in genomics, cancer and cancer genomics journals, as well as treatment-related consensus statements issued by professional societies. In ovarian cancer discourse, when the detection of a BRCA mutation suggests the necessity (and in some cases the inevitability) of surgery, the term ‘mutation’ functions essentially as a trope for disease. I trace this shift as it develops after the introduction of BRCA mutation screens, paying particular attention to the ways in which coextensive rhetorics make possible the meanings of mutation and the consequent substitute of genotype for phenotype in cancer management discourse. In particular, I focus on operative assumptions about risk, prevention and motherhood in order to make sense of how it is that women's reproductive organs are so easily and effectively diseased.

Embodying heredity: the new subject of risk

Well before the first BRCA gene was mapped, interest in a genetic predisposition for ovarian cancer was sparked by knowledge of the so-called ‘hereditary cancer syndromes’ affecting certain families. These syndromes explained why in some families, many women—especially young women—were diagnosed with breast and/or ovarian cancer. A study based on the Gilda Radner Familial Ovarian Cancer Registry from 1981 to 1991 established that ovarian cancer risk could indeed be understood as a genetic predisposition for some families.1 In these families, in which two or more first- and second-degree relatives had ovarian cancer, risk was apparently linked to an autosomal dominant trait (meaning one mutated copy of the gene confers a risk, although because of ‘variable penetrance’ not all women with the mutation will actually get cancer). For researchers studying families in the Radner Registry, mother–daughter and sister–sister cases were typical of the vast majority of cancer syndrome families (Piver et al., 1993a). The risk for women with a family history that included mother–daughter cases was estimated to be about 41% but could be as high as 50% (Piver et al., 1993b); the lifetime risk for women in the general population is 1.8% (see King et al., 2003, p. 643). Two years later, participants in a workshop sponsored by the US National Cancer Institute and National Center for Human Genome Research (now called the National Human Genome Research Institute) concluded that 1% of women with ovarian cancer have a family history consistent with an autosomal dominant mode of transmission, although an additional 7% of women with ovarian cancer with a family history of at least one affected first- or second-degree relative are thought to be at ‘moderate’ risk for the disease (Gallion & Park, 1995). ‘Ultimately,’ concluded the Radner Registry report, ‘early detection of the individual at risk for ovarian cancer will depend on finding the abnormal gene’ (Piver et al., 1993a, p. 588).

Talk of a gene afflicting cancer syndrome families began in the early 1990s with the theory of Mary-Claire King, a geneticist at the University of Washington and her colleagues, that a ‘breast cancer gene’ called BRCA, on the short arm of chromosome 17, was linked to heightened risk for the disease (Hall et al., 1990) and in 1994 the sequence of BRCA1 was published (Miki et al., 1994). In 1995, the breast cancer susceptibility gene ‘BRCA2’ was identified (Wooster et al., 1995) and by 2001 researchers had discovered nearly 1,000 mutations of the two genes. Clinical data demonstrate that mutations of the BRCA genes can increase risk for breast cancer anywhere from 30% to 80% by age 70 (Boyd, 2001).2

In 1991, soon after the first BRCA1 report, Narod et al. (1991). published the finding that ovarian cancer risk was also linked to mutations of this gene, and it is now known that mutations of BRCA2 are also implicated in ovarian cancer incidence. There have been several large studies of ovarian cancer demonstrating that mutations of BRCA1 and BRCA2 can elevate a woman's lifetime risk for the disease (e.g. Risch et al., 2001; Olopade & Artioli, 2004). According to clinical data, women with BRCA mutations face anywhere between a 10% to 60% chance of developing ovarian cancer depending on which mutation they inherit of either BRCA1 or BRCA2, with BRCA1 mutations believed to confer a higher risk and account for an earlier age of onset.

Treatment recommendations for high-risk women have often included oophorectomy (e.g. Piver et al., 1993a) largely because ovarian cancer is such a deadly disease. Ovarian cancer causes more deaths than any other cancer of the female reproductive system with the exception of breast cancer (Centers for Disease Control, 2005). In 2004, 26,000 cases were diagnosed and almost 16,000 of these women died (American Cancer Society, 2005). It is unclear just how many women have elected to undergo this procedure, but it is clear that as early as 1995, recommendations that women considered to be at exceptional risk undergo surgery were common. In 1995, a US National Institutes of Health Consensus Development Panel on Ovarian Cancer concluded that:

the probability of a hereditary ovarian cancer syndrome in a family pedigree increases with the number of affected relatives, with the number of affected generations, and with young age of onset of disease. There, prophylactic oophorectomy should be considered in these settings with careful weighing of the risks and benefits. The risk of ovarian cancer in women from families with hereditary ovarian cancer syndromes … is sufficiently high to recommend prophylactic oophorectomy in these women at 35 years of age or after childbearing is completed (Seltzer et al., 1995, p. 493).

Not all recommendations for prophylactic oophorectomy have been so strongly worded. In large part this is because prior to 2002, very little scientific evidence existed demonstrating the health benefits of oophorectomy. Particularly problematic was the fact that no prospective case control trials had been carried out, and of the studies that did show a benefit, the study subjects were not BRCA mutation carriers. A Consensus Statement by the Cancer Genetics Studies Consortium organized by the US National Human Genome Research Institute did not include an explicit recommendation for oophorectomy, writing that ‘[o]bservational data have so far failed to demonstrate statistically significant evidence for risk reduction’ (Burke et al., 1997, p. 1000). Much of the concern over oophorectomy was the result of researchers' observations that as many as 10% of women undergoing oophorectomy were diagnosed with ovarian-type cancers years later (e.g. Tobacman et al., 1982). These women were diagnosed with peritoneal carcinoma, a disease indistinguishable from ovarian cancer.3

Cautious language concerning prophylactic oophorectomy was quickly replaced with enthusiastic assessments after the publication of two prospective studies demonstrating that ovarian cancer risk can be lowered after oophorectomy for BRCA carriers (Kauff et al., 2002; Rebbeck et al., 2002). In the Kauff et al. study, after a mean follow-up of two years, one out of 98 women who had oophorectomy developed peritoneal cancer compared to five out of 72 women who had elected to undergo surveillance only (four ovarian cancers, one peritoneal cancer). In the Rebbeck et al. study, two out of 259 women undergoing oophorectomy were diagnosed with peritoneal cancer during the approximately eight years of follow-up as opposed to 58 out of 292 women who opted for surveillance and were later diagnosed with ovarian cancer. For the Kauff group, the results provided ‘strong support for including discussion of risk-reducing salpingo-oophorectomy as part of a preventive-oncology strategy for women with a BRCA1 or BRCA2 mutation’ (n.p.).4 For Rebbeck et al., ‘[o]n the basis of the results of our study, we advocate prophylactic oophorectomy to reduce the risk of ovarian and breast cancer in women with BRCA1 or BRCA2 mutations’ (n.p.).

These 2002 oophorectomy studies are widely cited in the cancer genetics literature as scientific evidence that the procedure is, on the whole, a good idea for women with BRCA mutations. Recommendations include salpingo-oophorectomy because of the link between BRCA mutations and cancer of the fallopian tubes. Writing in the Journal of Clinical Oncology in 2003, Douglas Levine and colleagues claim ‘risk-reducing salpingo-oophorectomy is currently recommended for BRCA mutation carriers after completion of childbearing or at age 35 years’ (Levine et al., 2003). A 2004 literature review published in the journal Gynecologic Oncology concluded ‘prophylactic surgery conveys the greatest risk reduction and is a procedure that needs to be offered to all BRCA1 and 2 germline carriers at this time’ (Rosen et al., 2004, p. 285). In a 2004 editorial in the Journal of Clinical Oncology, researchers suggested that given the available data regarding risk reduction from oophorectomy, the question for women with BRCA mutations is no longer ‘whether’ to have the procedure but ‘when’ (Garber & Hartman, 2004).

The discourse around oophorectomy suggests an understanding among researchers and clinicians that ovarian cancer constitutes a grave threat to women that warrants no less than extreme and irreversible medical interventions. To some extent, the recent attention to oophorectomy is an extension of earlier medical discourse regarding care for women at risk for familial cancer. The 2002 prospective studies simply provide concrete evidence regarding the efficacy of prophylactic oophorectomy in decreasing risk for ovarian cancer, especially for women with BRCA mutations.

I want to suggest, however, that the identifier ‘BRCA carrier’ is not merely another way of saying ‘woman at high risk’ or, as cancer syndrome family researchers put it, ‘woman with a family history of familial ovarian cancer’. Rather, BRCA mutation status is a qualitatively, epistemologically unique risk category; when cancer geneticists linked BRCA mutations to inherited ovarian cancer, a new ‘risk subject’ emerged. By de-linking BRCA risk from familial risk, the body is marked in different, perhaps unexpected ways, the implication being that more women will be affected by cancer genomics. Indeed, it is the marking of bodies that suggests BRCA mutations are materially different from risk ascertained from family pedigree studies, as risk can be ‘seen’ after a genetic test. Genetic tests have become, for physicians and oncologists, a form of surveillance, substituting traditional, yet somewhat unreliable methods such as sonography and CA-125 testing. Genetic screens do not detect tumour development—they screen for risk; it was the expressed hope of researchers, in fact, that the promise of genetics research was no less than the development of screens to detect early molecular changes preceding tumour development (Bourne et al., 1991; van Nagell, 1991). Yet the interventions have not changed suggesting that a detected BRCA mutation operates as a trope, or shorthand, for disease in ways that inferred risk from family studies did not.

Prior to the development of BRCA screens, researchers estimated that about 1% of women with ovarian cancer have a family history consistent with autosomal dominant trait transmission. Data emerging from subsequent studies of women with ovarian cancer revealed that many more women with the disease test positive for mutations—as many as 10% (Haber, 2002). So although BRCA mutations likely explain the vast majority of inherited ovarian cancer (according to one report, nearly 90%; see Boyd, 2001), they are perhaps much more common in women with cancer and play a role in cancers unrelated to family cancer syndromes.

In fact, family history is now neither necessary nor sufficient in predicting the presence of BRCA mutations and their role in ovarian cancer. BRCA mutations are increasingly de-linked from consideration of family history, thus allowing them to be categorized as a distinct risk factor. With the design of population-based studies (and here I mean the general population, not population genetics) participants are not chosen because of a family history for the disease; they are chosen only if they have a personal history of ovarian cancer. In one study, researchers observed that the majority of women with ovarian cancer did not report a significant family history of the disease (Cass et al., 2003). Although the researchers relied on self-reporting of family history—a potentially significant limitation of the study—the Cass study and others like it nevertheless point to a tendency among cancer geneticists to de-link BRCA-related cancers from cases associated with the cancer syndromes. As one researcher put it, family history alone is a ‘poor predictor’ (Colgan et al., 2001). As a result, a woman—whether she has had cancer or not—no longer needs to be a member of a cancer syndrome family to be a candidate for genetic testing.

Other studies have shown just what some extra-familial variables might be. In 2001, a large study of women with ovarian cancer revealed that women with non-mucinous tumours (those that are poorly differentiated and much harder to treat) were significantly more likely to test positive for BRCA mutations; researchers concluded that it is ‘reasonable to offer genetic testing to all women with invasive nonmucinous ovarian cancer’ (Risch et al., 2001, p. 708). Ethnicity has also emerged as a marker for genetic risk. In one study of Jewish women with ovarian cancer researchers concluded that ‘[i]n light of the high probability of BRCA mutations in Jewish patients with ovarian carcinoma, we currently discuss genetic testing with all newly diagnosed, Jewish patients with ovarian/peritoneal or fallopian tube carcinoma’ (Cass et al., 2003). Histology and ancestry have thus become evidence of inherited risk, evidence that does not necessarily appear in conjunction with family history of the disease. Even age of onset is no longer a reliable predictor of a mutation, as women with BRCA2 mutations do not develop ovarian cancer at an appreciably younger age than women in the general population.5

Many variables, then, regardless of family history or age of onset, are grounds for conducting genetic screens. For women with ovarian cancer, genetic tests are potentially valuable, as the information they reveal may affect treatment options—women with these mutations may respond differently to medication and some are even expected to have better or worse chances for surviving the cancer (Rubin, 2003). How their risk is retroactively defined may, however, have implications for family members. When a woman tests positive for a mutation, family members—even those individuals whose family would not meet the definition of a cancer syndrome family—risk feeling pressure to be tested. Although test results are typically not disclosed without the consent of the woman tested, many clinicians feel that women have a responsibility to tell family members who may need to avail themselves of risk-prevention strategies such as prophylactic surgery (e.g. Cass et al., 2003). Thus, although the number of women suspected to test positive for a BRCA mutation remains comparatively low (5–10% of ovarian cancers are thought to be hereditary), of those women, the number that will be tested may increase. Once tested and marked ‘at risk’, these women become targets of interventions such as prophylactic surgery.

The definition of the risk subject is thus broadened considerably as a result of these developments. Yet the very meaning of risk is also undergoing significant transformation. As I suggested above, the information revealed by genetic screens is taken to be materially different from risk estimates derived from family histories. In the early 1990s, researchers suggested that discovery of the BRCA genes and the development of reliable tests would permit the identification of women for whom family history had in fact resulted in the transmission of a suspected cancer susceptibility trait—women ‘truly’ at risk. BRCA tests, some claim, reveal ‘actual risk’ of cancer (Spear et al., 1999) of ‘genetically-defined’ (Eisen & Weber, 1998, p. 798) women. Writing in the Journal of Clinical Oncology, researchers from cancer centres in the United States and Canada concluded that ‘[t]he strongest risk factor for ovarian cancer is the presence of an inherited mutation in 1 of the 2 ovarian cancer susceptibility genes, BRCA1 or BRCA2’ (Liede et al., 2002).

Discourse about oophorectomy clearly shows that the information revealed through BRCA tests is materially different—genetic markers and the molecular basis of heredity that they reveal is information that can, with confidence, provide the grounds for medical intervention. As I suggested earlier, a consensus is clearly emerging that all women with BRCA mutations should consider prophylactic oophorectomy. As more and more articles appear on the pages of medical journals suggesting that all BRCA carriers consider surgery (some even assert that this is already the case and that most women considered to be high risk do in fact opt for this surgery), the more likely it will be that physicians and oncologists will discuss this type of treatment for risk. As knowledge of the relationship between the BRCA genes and breast cancer is already widespread among physicians, specialists and patients alike, it is likely that awareness of the link to ovarian cancer will also spread.

Here I want to call attention to how a BRCA mutation comes to operate as a trope for disease, as the treatment recommendations for healthy women at risk for ovarian cancer because of the presence of a BRCA mutation suggest that, in this discourse, risk is the disease, a slippage that is theoretically and rhetorically possible because of the implicit importance in our culture attributed to heredity. Arguably, researchers equated heredity with disease when they recommended prophylactic oophorectomy well before the commercialization of BRCA screens and well before evidence existed that the procedure would in fact significantly reduce risk. The BRCA screen, however, marks risk in materially different ways suggesting that the information these tests reveal provides a stronger justification for extreme measures.

Of the members of cancer syndrome families, for the vast majority, their risk justified intense surveillance, not prophylactic surgery. The BRCA screen, by contrast, reveals putatively ‘real’ risk, replacing traditional screening methods such as sonography and CA-125 serum testing, the latter two used to detect ovarian cancer. BRCA screens have become, in the words of cancer researchers, substitutes for screening; they are a method of ‘primary’ prevention since the goal is not detecting early-stage tumours but detecting risk. The substitution of BRCA screens for cancer surveillance, or ‘secondary’ prevention, is further enabled by the routine dismissal of the latter's potential effectiveness. In 10 articles in which researchers discount the efficacy of screening, not one cited a 1991 study about the effectiveness of sonography, especially for women with a family history of the disease.6 At best, they characterize the extant data as mixed; at worst, they claim that ‘no screening for ovarian cancer has yet been proved effective for women in any risk category’ (Kauff & Barakat, 2004, p. 277, emphasis added). Yet in 1991, some researchers were reasonably optimistic about the benefits of screening women at high risk:

TVS [transvaginal ultrasonography] is particularly effective as a screening method in women whose primary or secondary relatives have documented ovarian cancer. By limiting screenees to those with a positive family history of ovarian cancer, the authors found that the prevalence of ovarian cancer in the screened population increased 10-fold. As a result, the positive predictive value of TVS increased proportionately. Three primary ovarian cancers were detected in 776 asymptomatic women. All three women had Stage I disease and all are now alive and well (van Nagell, 1991, p. 91).

Conceivably, use of TVS in the above study presumed that hereditary ovarian cancer is a variably penetrant disease necessitating inclusion of family history and a cautious, less intrusive approach to prevention. In contrast, BRCA mutations necessitate no less than surgery—betraying the belief among researchers that risk is functionally 100%. Developments in ovarian cancer detection and treatment, such as more and more calls for genetic testing and strongly worded recommendations for aggressive, sometimes experimental, procedures for healthy women, suggest that in the practice of ovarian cancer research and treatment, heredity is the disease. Just as a doctor would not recommend surveillance instead of treatment for a woman with cancer, so too surveillance for a woman ‘at risk’ is not a rational choice in the evolving discourse of ovarian cancer.

By all accounts, hereditary ovarian cancer is variably penetrant—risk from BRCA mutations ranges from 10% to 60% depending on the mutation, where the mutation is located along the gene (Risch et al., 2001) and a whole host of environmental factors (King et al., 2003). This is perhaps why BRCA mutations are found in a larger percentage of ovarian cancer cases than previously estimated. Ideally, recommendations would take this into account—just as pre-BRCA recommendations for women with a family history considered which women had ovarian cancer (e.g. first degree relative vs. second degree relative) in order to determine who was at ‘high’ or ‘moderate’ risk and tailor treatment recommendations accordingly, prophylactic oophorectomy is all too often recommended for all BRCA carriers, regardless of any mitigating (or aggravating) variables. This is especially problematic as clinical data are limited to the study of select mutations (recall that there are hundreds). Risk assessment, then, is often performed by determining whether the chemical makeup of the mutation suggests that it is ‘deleterious’,7 not by whether clinical data demonstrate the approximate risk it confers.

Prophylactic oophorectomy was not even taken very seriously prior to the sequencing of the BRCA genes, suggesting that researchers were waiting until ‘real’ evidence of risk could be ascertained through genetic testing before studying it closely. Even before the commercialization of BRCA tests, researchers suggested that information about risk would result in an increased interest in prophylactic oophorectomy, thus necessitating more and better designed studies to determine its efficacy in reducing that risk (Struewing et al., 1995). Wrote one physician: ‘The identification of the BRCA genes and the availability of genetic testing for BRCA mutations have underscored the need for improvements in early detection and prevention of breast and ovarian cancer’ (Haber, 2002). Indeed, the 2002 prospective studies of prophylactic surgery were carried out with the explicit purpose of determining whether or not BRCA mutation carriers would benefit—they have further fuelled calls for more genetic tests, especially for women with a family history of the disease (Haber, 2002).

My point here is not to dispute that some women are at risk and that they can reduce that risk by having their ovaries removed. Rather, it is to show that women's bodies, via discourses of risk, are represented in qualitatively different ways and are thus ‘accessible’ to medical practitioners in unique ways—a departure from the ‘genomics is progress’ paradigm that dominates public discourse about this science and its impact on medicine. The interventions that new representations of risk enable is a topic to which I now turn.

Embodying heredity: the treatment of risk

I have suggested that there is a materiality to BRCA mutations that is qualitatively, substantively, different from risk ascertained from family pedigrees. Family pedigree analysis requires knowledge of cancer histories and counsellors determine from those histories whether patterns suggest a hereditary element to cancer risk. Risk is based on information; it is itself information, and as such resides within the realm of the abstract. Genetic screens provide information that is used to conduct risk assessments, yet the body of the woman being tested is directly involved—in this case the tissue sample used to analyse her DNA. The screen provides material, bodily evidence of risk.

When geneticists, physicians and oncologists talk about the importance of BRCA screens, their material significance is acknowledged in claims that BRCA tests reveal ‘actual’ risk. Yet those measures do not, perforce, entail altering a woman's DNA in order to ameliorate the risk of BRCA mutations (and we are a long way from gene therapy of this sort, if in fact it will be possible at all). The BRCA literature is replete with statements regarding the importance of empowering women with the information they need to seek risk reduction ‘measures’. Thus the significance of the genetic screen is not that it allows the physician to alter the materiality of risk that it reveals; rather, its use value as information is what is significant—it merely signifies risk, it is not the embodiment of risk that can then be manipulated. What will be manipulated is a woman's body, in this case her reproductive organs. Reproductive organs are, in the final analysis, the embodiment of hereditary risk and as such become the site of intervention.

The relationship between representations of risk and its subsequent embodiment necessitates sustained feminist analysis. The question that invariably occupies bioethicists is whether women will be free to utilize the information revealed by genetic screens. BRCA tests are simply one more diagnostic tool ‘adding’ to our knowledge of breast cancer. Alternatively, following Barbara Katz Rothman (1986), feminists should think critically about what is meant by ‘information’ in genomics and what happens when that information is circulated and appropriated in ways that are historically unique and socially determined. Feminist analysis should include an examination of the relationship between representations of risk—the construction of what I am calling the risk subject and the representations of women's bodies this necessarily entails—and the interventions that these representations enable. I have thus shown how the various threads of inherited ovarian cancer discourse collectively enable the formation of a new subject of risk, one that is marked in ways far beyond the somewhat circumscribed categories of familial risk.

Of principal concern to this writer is the fact that women with BRCA mutations are subject to the same interventions as women identified as high-risk prior to BRCA research, which is to say, the same interventions as women with cancer. As I have already discussed, the treatment of choice is prophylactic oophorectomy, specifically salpingo-oophorectomy involving removal of both the ovaries and the fallopian tubes.

Prophylactic surgery is not a treatment to be taken lightly; if there was any hesitancy before among specialists, it was because absent a genetic test, one could not be sure that a woman had in fact inherited a mutation or even if a mutation explained her family's unusually high number of cancer cases. BRCA research, however, imbues the procedure with a degree of legitimacy that it previously lacked. We can now identify women ‘truly’ at risk and justify the same treatments prescribed for women with cancer. Treatments effectively make risk and disease one and the same.

The facile reclassification of ovaries from ‘at-risk’ to ‘diseased’ is possible in part, then, because of implicit assumptions about heredity in discourse about BRCA genes. The diseasing of ovaries and the subsequent call for their removal is further enabled, however, by habits of mind governing the importance of ovaries to a woman's happiness and well-being. The embodiment of medical discourse is possible only if the body is materially ‘available’ for intervention. Here representations of risk intersect with representations of women's reproductive organs, in this case, their ovaries.

The routine, cavalier removal of women's reproductive organs in the name of disease prevention marks the history of dominant, arguably misogynist, discourses of health, disease and the body. For example, the Halsted radical mastectomy, a highly disfiguring procedure involving removal of breasts, lymph nodes and chest wall muscles, was popular among cancer surgeons well after research showed that alternatives were equally effective for saving women's lives (see Lerner, 2001).8 Breast cancer ‘prevention’ has included even more extreme measures, such as oophorectomy and adrenalectomy in order to control a woman's oestrogen production (Huggins & Dao, 1953).9 And as recently as 2000, hysterectomy was the second most frequently performed major surgery in the United States. The majority of these operations, critics have argued, are unnecessary (Broder et al., 2000, p. 199).10 Oophorectomy has been generally regarded in a somewhat cavalier fashion. As recently as 1999, officials from the New York State Department of Health wrote that ‘[i]f a patient at risk for ovarian cancer is having abdominal/pelvic surgery, a prophylactic oophorectomy should be performed. It has been reported that elective routine removal of ovaries after age 40 would eliminate approximately 12–14% of all ovarian cancers’ (Runowicz, 1999, emphasis added). This recommendation is commensurate with surveys showing that a majority of physicians view routine oophorectomy as an appropriate method of cancer prevention.11 Recall that prescriptions regarding oophorectomy have often entailed the scorched-earth logic of radical mastectomy: in one study, researchers recommended that the oophorectomy be combined with hysterectomy and fallopian tube removal in order to protect women at risk for post-operative peritoneal cancer (see Tobacman et al., 1982; Karlan, 2004, p. 520).12

Curiously, the medical community recognizes that surgery alone in no way guarantees that a woman's cancer will be ‘cured’. And largely as a result of patient activism, breast cancer researchers recognize the appropriateness of lumpectomy in saving women's lives without exposing them to unnecessary surgery and disfigurement. Also largely the result of women's activism, hysterectomies continue to be scrutinized and alternatives explored for women with serious menstrual disorders.

Yet many of these same procedures continue to capture the attention of cancer researchers, not so much for women with cancer, as for women at risk for it. For healthy women with BRCA mutations, complete, double mastectomy is often touted in the cancer literature as a life-saving preventive measure, even though women with cancer are reasonably expected to consider lumpectomy instead. Mastectomy for BRCA carriers has been roundly criticized by breast cancer organizations,13 especially in view of the fact that residual breast tissue often remains and can become cancerous (Hubbard, 1996; Healy, 1997; Fasouliotis & Schenker, 2000). Although a study published in 1999 showed that women can reduce their risk dramatically with this procedure (Hartmann et al., 1999), ‘its efficacy in reducing this risk is uncertain, as cancer after prophylactic mastectomy has been well documented in the literature, with results varying from 1 to 19 percent’ (Fasouliotis & Schenker, 2000).

Like mastectomy, oophorectomy does not completely eliminate risk for cancer. The Kauff et al. and Rebbeck et al. studies published in 2002 did show a dramatic decrease in ovarian cancer; for some women, however, the procedure did not decrease risk for peritoneal cancer, a phenomenon observed for some time, suggesting what Eisen and Weber (1998) have called BRCA's ‘field’ effect. In a third study published in 2002, researchers suggested that ‘[t]he majority of the ovarian-type cancers observed in our cohort were determined to be of peritoneal origin and, we believe, would not have been prevented through prophylactic salpingo-oophorectomy’ (Liede et al., 2002, n.p.).

Unlike mastectomy, however, oophorectomy increases risk for other diseases; it also impacts on the reproductive choices available to women. And since the 2002 Rebbeck et al. and Kauff et al. studies demonstrated a significant reduction in breast cancer risk, we can expect that the procedure will be prescribed for BRCA carriers at high risk for breast cancer. Several researchers have speculated that women will in fact prefer oophorectomy to mastectomy to avoid the disfiguring consequences of the latter (Risch et al., 2001). Concluded one physician, ‘[t]he effectiveness of prophylactic oophorectomy in carriers of BRCA mutations provides a strong rationale for genetic testing in women with a strong family history of breast cancer’ (Haber, 2002).

Oophorectomy further increases risk for diseases other than cancer. Ovaries are fundamentally important to a fully functioning endocrine system, and as such, their removal impacts on women in ways that mastectomy does not. It is widely known, for instance, that menopause increases the risk of heart disease. According to at least one report, oophorectomy resulted in a substantial increase in coronary heart disease for post-menopausal women, a risk not seen in women undergoing natural menopause (Colditz et al., 1987). The sudden decrease in oestrogen levels may, therefore, result in more negative health effects than the more gradual decline as a result of natural menopause. This danger extends to pre-menopausal women (who are likely to have prophylactic surgery to reduce risk): according to the 1995 National Institutes of Health consensus statement on oophorectomy, the procedure ‘may result in a significant reduction in life expectancy due to cardiovascular disease and osteoporosis in pre-menopausal women who have bilateral oophorectomy, compared with women with retained ovaries’ (Seltzer et al., 1995, p. 493; see also Aitken et al., 1973).

The results of the recently completed Women's Health Initiative call into question whether hormone replacement therapy (HRT) can help prevent some of the side effects of menopause such as heart disease and osteoporosis. Even if women consider hormone replacement therapy in order to compensate for the health effects of premature menopause, BRCA mutation carriers are at risk for breast cancer as well and may be reluctant to take combined oestrogen–progestin HRT (Karlan, 2004, p. 520; although one researcher concluded that HRT merely negates the reduction in breast cancer risk that oophorectomy provides). The only other option for women who choose not to take combined HRT is to take oestrogen-only HRT, thus placing themselves at higher risk for uterine cancer. This conundrum has led researchers and practising physicians to recommend that these women undergo hysterectomy at the time of salpingo-oophorectomy, which may require major surgery. New methods such as ‘laparoscopic-assisted vaginal hysterectomy’ (in which the uterus is removed through the vagina), presents its own complications. According to one team of researchers, ‘the single biggest predictor of complications associated with laparoscopic procedures was the inclusion of hysterectomy as part of the procedure’ (Kauff & Barakat, 2004, pp. 277–8). (Oophorectomy carries its own risks: in one study, four women out of 98 experienced ‘complications’ such as infected wounds, punctured bladders, punctured uteri, and bowel problems: Kauff et al. (2002)).

As one could imagine, the decision by a healthy woman to undergo prophylactic oophorectomy is a difficult one. She must calculate, on the one hand, the risk associated with her BRCA mutation and how much an oophorectomy will reduce that risk. On the other hand, a woman must calculate the risk of getting cancer after surgery, the risks associated with the procedure itself, and the health risks associated with premature menopause. In order to aid the decision-making process, ‘decision analysis’ experts have developed models for calculating the relative benefit of oophorectomy.

Models, however, are only as sound as the assumptions that inform them. For instance, the decision model developed by Grann et al. (2002) assumes that tamoxifen can substantially reduce risk for breast cancer, but the evidence is mixed as to whether women with BRCA2 mutations would actually benefit from this form of chemoprevention (King et al., 2001). The model also assumes that HRT does not increase breast cancer risk. Grann et al. concluded that since women with BRCA mutations are already at a very high risk for breast cancer, HRT impact is negligible and thus should not be part of the risk analysis.14 This argument, however, makes sense only for women who have an 85% lifetime risk for breast cancer. BRCA-linked breast cancer risk is in fact highly variable and ranges from 40% to 85% depending on the mutation that a woman has—and many women test positive for mutations that have not been studied in clinical trials. The Grann decision analysis model further assumes that the risk for heart disease is the same as for women in the general population, despite evidence suggesting that the risk may be higher for women experiencing surgical menopause. And finally, the model assumes that all women face the same risk for mortality from ovarian cancer even though some studies suggest that women with BRCA mutations have a better chance of survival than women without (Cass et al., 2003).

Even with the questionable assumptions informing the Grann model, the greatest life expectancy gain predicted is 4.6 years if a 30-year-old woman undergoes oophorectomy and chemoprevention (tamoxifen). A different model predicts more modest gains: 0.3 to 1.7 years for a 30-year-old woman who chooses oophorectomy (Schrag et al., 1997).

Although researchers do recognize the many health benefits of intact ovaries, discourse about women at risk for BRCA-related cancer significantly waters down these benefits, nearly to the point of rendering them irrelevant to any treatment decision calculus. Indeed, the health benefits of intact ovaries are rarely considered substantial enough to weigh against the risk benefit associated with their removal. Thus, the short- and long-term side effects of oophorectomy often recede, rhetorically, in a discourse in which the telos is the reduction of cancer risk, seemingly at all costs. In one crucial study supporting the efficacy of oophorectomy, the researchers first construct a decision calculus that clearly downplays the side effects associated with oophorectomy:

The primary negative consequence of prophylactic oophorectomy in premenopausal women is premature menopause, which may be associated with increased risks of osteoporosis and cardiovascular disease. Hot flashes, vaginal dryness, sexual dysfunction, sleep disturbances, and cognitive changes associated with menopause may affect the quality of life. However, the risk is balanced by the morbidity and mortality associated with breast and ovarian cancer in carriers of BRCA1 or BRCA2 mutations, and these symptoms may be managed by hormonal or nonhormonal medications (Rebbeck et al., 2002, emphasis added).

After strategically qualifying cancer risks with an explicit reference to mortality, the researchers conclude by recommending oophorectomy:

On the basis of our study, we advocate prophylactic oophorectomy to reduce the risk of ovarian and breast cancer in women with BRCA1 or BRCA2 mutations. In deciding whether to undergo the procedure, a woman should take into account how long she wishes to maintain fertility, and she should receive counseling about the risk and benefits of prophylactic oophorectomy. The decision should also be made with the knowledge that current surveillance regimens have not been shown to affect the incidence of late-stage ovarian cancer. Although opinion is divided on the use of hormone-replacement therapy after prophylactic oophorectomy, the decision to use estrogens should be based on a consideration of symptoms that affect future health and the quality of life. Some centers routinely recommend hormone-replacement therapy after prophylactic oophorectomy until the age of 50 years, and many women consider prophylactic oophorectomy unacceptable without this option (Rebbeck et al., 2002, emphasis added).

These statements reveal the implicit belief among researchers that a woman's fate is, after all, in her genes, and with newfound knowledge of BRCA genes, heredity is materially observable in ways that justify all feasible interventions. Other forms of risk simply do not count for much (for instance risk for heart disease)—BRCA mutations have come to stand in, it seems, for disease itself. These statements further reveal that beyond the considerable cache that genetics carries in this culture, the legitimacy of oophorectomy is also possible because of the metonymical reduction of ovaries to their reproductive function. When cancer researchers effectively dismiss the health importance of ovaries for the woman at risk for cancer, they betray the belief that ovaries provide women with the ability to procreate and nothing else; for women who no longer need or want ovaries for this purpose, they are dispensable.

Reproduction and cancer risk have a lengthy, shared history, as the two are often inextricably linked. According to the ‘oestrogen hypothesis’, the more oestrogen a woman is exposed to in her lifetime, the higher the risk for cancer. Parity (the number of childbirths) can increase or decrease breast cancer risk, since oestrogen levels drop considerably when a woman is pregnant. Early parity has an even greater protective effect, not only because of the shorter duration of oestrogen exposure over the course of a woman's reproductive years, but because breast cells differentiate earlier and thus become less susceptible to mutagens. Although not related to the oestrogen hypothesis, parity can decrease risk for ovarian cancer since pregnancy reduces the number of ovulation cycles and by consequence reduces the chances that ovarian cells will mutate. It is beyond the scope of this article to discuss the limits of the parity argument,15 but whether or not it is valid, it is still the case that it is a risk factor that has received a great deal of attention by the cancer research community. Indeed, whenever one comes across a list of typical risk factors for breast cancer, it is virtually always mentioned. My point here is that the relationship between oestrogen, ovulation and cancer is a complicated one, and that whether or not parity is a risk factor, the perceived relationship between reproductive history and cancer risk influences the nature of advice and treatment that women receive from their doctors.

In cancer discourse, a number of variables act as proxies for reproductive history, such as socioeconomic status, religion and sexual orientation. As a result, parity has been implicated in higher rates of breast cancer among lesbians, nuns and women living on Long Island. There is, then, a way in which women's lifestyles are indirectly pathologized in cancer discourse. No one of course suggests that women have as many children as possible and as early as possible in order to lower risk for cancer; the unspoken (although sometimes explicit and very reasonable) assumption is that women should be free to choose how many children they have and when and that cancer prevention should entail the development of more effective surveillance and treatment.

However, for women at risk for ovarian cancer, reproduction is an integral part of prevention discourse, since oophorectomy results in infertility. Cancer researchers have to confront how cancer risk management impacts on the reproductive choices available to women. Authors of the research reports I examined are careful to say that women at risk for ovarian cancer should undergo prophylactic oophorectomy as soon as possible after childbearing is complete (e.g. Rebbeck et al., 2002). In the effort to protect reproductive autonomy, researchers implicitly acknowledge that fear of cancer should not interfere with other concerns, in this case motherhood. Motherhood is, in fact, the only exception to the norm—in my research thus far, I have yet to find a scientific report that acknowledges the acceptability of postponing, or avoiding altogether, oophorectomy because of health concerns, a somewhat surprising outcome since age of onset varies for each BRCA carrier. A BRCA1 carrier may want to undergo oophorectomy sooner than later as her diagnosis will likely occur earlier than diagnoses for women in the general population. Alternatively, a BRCA2 carrier, who will likely be diagnosed at the same age as women in the general population, may want to postpone (or avoid altogether) oophorectomy in order to reduce risk for heart disease and osteoporosis.

This discourse reveals the shared belief among researchers that reproductive autonomy must be taken seriously when developing cancer prevention guidelines. For women both able and desirous of motherhood, cancer risk management can and should be balanced against the decision to start a family. As such, it apparently addresses the concerns of bioethicists who would argue that genomics technology must be critically evaluated to determine whether it serves women's interests. Those interests, however, are narrowly defined. If recommendations were in fact devised with feminist concerns in mind, one would expect there to be a serious discussion about the negative consequences of premature menopause beyond infertility, a discussion that presumes that a woman's happiness and well-being depends on many factors.

The near singular focus on reproductive potential when evaluating the benefits of prophylactic oophorectomy is by no means unique to this discourse; it reflects, rather, widely held assumptions about women and reproductive capacity. Women are often hailed as mothers or potential mothers in the struggle over abortion rights, debates about prenatal genetic testing and attempts to regulate their employment options.16

Motherhood, whether potential or realized, is constantly negotiated and regulated. The timing of oophorectomy is a central feature of discourse about inherited ovarian cancer, suggesting that motherhood is similarly disciplined even if in circuitous, unsuspecting ways. Reproductive rights, so it seems, cannot be fully exercised without incurring some health risks. A consensus exists among researches and practising oncologists that women should undergo oophorectomy at age 35 at the latest; childbearing, then, ought to occur relatively early in life. Fertility trends show that women are having children later in life, often well into their 30s and early 40s (Aliyu et al., 2005). Among cancer researchers, however, 35 is the border between young and old pregnancy, reflecting the norm in obstetrics that define 35 as advanced as ‘advanced maternal age’ (Heffner, 2004). Oophorectomy after age 35 is considered risky and unadvisable, a concern not only for women who have not been able to or wanted to have children before that point, but for women who will experience potentially significant health effects from experimenting menopause nearly 20 years earlier than most women.

For women who postpone oophorectomy beyond age 35 in order to conceive, however, it is just as likely that they will encounter stiff resistance from their medical advisers. At least one report has shown that women with BRCA1 mutations tend to be diagnosed six to nine years earlier than women with BRCA2 mutations (Cass et al., 2003). ‘This finding has significant implications,’ the authors of the study argued, ‘for counseling BRCA mutation carriers regarding the timing of preventative interventions’ (Cass et al., 2003).17 The upshot, then, is that from the perspective of cancer researchers, the acceptability of postponing oophorectomy for childbearing is conditional. For BRCA1 mutation carriers, at least, early motherhood is necessary in order to reap the benefits of prophylactic surgery. Women for whom motherhood is not possible by age 35 or younger, as a result, face greater cancer risks:

With regard to the question of timing, there is uniform consensus, for obvious reasons, that risk-reducing salpingo-oophorectomy should be deferred until childbearing is complete. However, as more of our patients delay child bearing into their late 30s and 40s, BRCA1 mutation carriers, in particular, begin to expose themselves to a nontrivial rate of ovarian cancer (11–21% by age 50). Although carriers of BRCA2 mutations are at less risk of ovarian cancer during the reproductive years, mutation carriers who defer oophorectomy until the 40s and beyond may lose the substantial protective effect of salpingo-oophorectomy against breast cancer (Kauff & Barakat, 2004, p. 277, emphasis added).18

In this passage, researchers come close to suggesting explicitly that women change their reproductive behaviour in order to deal with their elevated cancer risk. There are of course many reasons why women delay childbearing, including but not limited to pursuing an education, a career, or having to wait for financial reasons. These choices, whether or not of their own making, increase women's risk for cancer, a seemingly irrational path to take given the ‘substantial protective effect’ of oophorectomy. This bears a great similarity to turn-of-the-century theories about the impact of reproductive organs on a woman's well-being. Physicians believed that a woman's uterus—present exclusively for the propagation of the species—could in turn be the source of a range of illnesses if a woman did not procreate (Ehrenreich & English, 1978). Drawing particular scorn were women who delayed childbearing in order to pursue an education and economic independence from men.

Women at risk for inherited ovarian cancer, then, are implicitly rewarded for motherhood, but only young motherhood. The subject of this discourse is not racially or socially marked, yet the association of young motherhood with better health likely contradicts the larger, public discourse about the appropriateness or desirability of such a status. Single women, for example, are discouraged from starting families, largely because we do not provide women with the means for supporting them. African-American motherhood is almost universally devalued, even pathologized. As legal studies scholar Dorothy Roberts (1999) has documented, social critics typically direct their ire towards black women—not institutional racism—when trying to explain phenomena such as poverty and incarceration. Young black women, despite higher than average fertility rates nevertheless suffer greater rates of breast cancer incidence and mortality, a statistic that flies in the face of the widely accepted oestrogen hypothesis and provides a dangerous point of articulation to the larger discourse that Roberts and others have documented. With regard to inherited ovarian cancer, feminists need to be particularly concerned with whether or not black women, especially young black women, are disproportionately encouraged to undergo oophorectomy.

Conclusion

In this article I have discussed the ways in which genetics research both changes how risk is defined and how it is treated in cancer research. I have also suggested that the legacy of gender bias in medicine over-determines the specific impact of the genomics revolution on women's lives. As BRCA research expands and evolves, I argue, more women will be considered ‘high-risk’ and their bodies ‘diseased’—developments that necessitate feminist scrutiny of the intersections of gender ideology, science and medicine.

Based upon the research and analysis presented thus far, I believe genetic tests must be re-evaluated as an acceptable part of the routine care for women at risk for BRCA-related cancers, although long-term genetics research may very well result in the development of effective, less devastating methods for detecting and treating these diseases.19 Like other feminist critics of science, I want to suggest alternatives to the medical research and clinical practice discussed thus far. However, I want to approach the subject of an alternative in a somewhat novel way: rather than describe what I suppose feminist BRCA research would look like, I want to suggest that a research programme investigating ovarian cancer as an environmental health problem would be an effective way of critically evaluating genomics from a progressive, materialist-feminist perspective. What are the contradictions between the discourse of heredity and the lived experiences of women? And how might feminists exploit those contradictions in order to challenge the discourse of heredity and its potential embodiment? As Emily Martin argues, embodiment provides a site of experiencing such contradictions and articulating alterative discourses and politics.

I have already mentioned that some geneticists openly acknowledge the role of environmental carcinogen exposure in explaining the variable ‘penetrance’ of BRCA mutations. According to King et al. (2003), risk for ovarian and breast cancer appears to increase with age for BRCA mutations carriers, much like for women in the general population. Age-related risk may very well have something to do with exposure to carcinogens that make it more likely that a mutated BRCA carrier's one good copy of the gene will mutate and thus stop functioning as a tumour suppressor.

For environmental epidemiologists, the connection between chemical pollution and cancer incidence makes sense. Researchers have observed, for instance, the link between ovarian tumours in wildlife and pollution of habitat, particularly when herbicides are present (Steingraber, 1998). For BRCA mutation carriers, extending this research to the study of human cancers may very well serve their interests as much as genetic research proper. Already, research about the so-called ‘oestrogenic’ synthetic chemicals called organochlorines has resulted in provocative claims that the historical rise in breast cancer incidence may be due the introduction of these chemicals into industrial and agricultural production.20 The rise in breast cancer cases, like many other cancers (especially of the fatty tissues) cannot be explained by the so-called ‘well-established’ risk factors: taken together, all of the risk factors for breast cancer, including inherited risk, explain only half of all breast cancer incidence in the United States.21

In the case of ovarian cancer, the potential link between environmental carcinogens exposure and incidence of the disease is virtually unexplored while biomedicine pursues genetic interventions that will very likely take several decades to come to fruition, if at all. This tension, and ultimate trade-off, between the genomics model of disease and the environmental health model is the result of the competing political perspectives they both represent and help articulate, if only indirectly. If one assumes, as I do, that disease is a marker of social relations, as well as a marker of biological malfunction, disease prevention can take the form of post hoc medical intervention and/or social policy that transforms the conditions of peoples' lives. That genomics privileges the former over the latter indicates its investment in existing social arrangements. Put simply, the gene replaces institutions and social arrangements as the cause of disease, often meaning that the victims of those institutions and social arrangements are ‘blamed’ for being sick. For women at risk for ovarian cancer, this further involves impugning their reproductive choices rather than the social and political environment in which they make them.

The alternative, I want to suggest, is the appropriation of progressive public health language when explaining and developing strategies for the prevention of women's cancers. Practically speaking, this would entail a well-funded investigation into the connection between environmental pollution and women's cancers. Recognizing that causation is difficult if not impossible to conclusively establish with the scientific tools already available, it would be necessary in the short term to embrace the ‘precautionary principle’ and enact strict environmental regulations for the purpose of dramatically reducing the use and release of synthetic chemicals.22

Leaving aside the question as to whether a materialist-feminist genetics research programme would be materially different from the genetics research we have today, for now I want only to suggest that the research/regulatory alternative discussed here performs the analytic and political work necessary for a socialist-feminist critique of genomics—first by recognizing that genomics performs significant ‘ideological work’ for economic and political elites by providing novel ways of avoiding any serious consideration of the relationship between health and social relations, and second, by recognizing that the impact of genomics on women's bodies and lives cannot be wholly explained by the relationship between genomics and a for-profit, pro-corporate medical industrial complex. Rather, feminists must fully account for why women in particular pay a high price for the choices research and policy elites make when it comes to medicine and public health policy.

Notes

1. Gilda Radner, a victim of ovarian cancer, was a famous American actress and comedienne.

2. We all inherit two copies of the BRCA genes at birth; according to the ‘two-hit hypothesis’ of cancer onset, both copies of a so-called susceptibility gene have to mutate before cancer can develop. So, women with two working copies of a cancer susceptibility like BRCA have a lower risk than women who inherit a damaged copy; they get cancer when the one and only working allele mutates—something quite likely to happen due to the vagaries of normal cell division and the panoply of environmental mutagens to which women are exposed on a daily basis.

3. This cancer is called ‘primary peritoneal carcinoma’ or ‘papillary serous carcinoma of the peritoneum’, or simply PSCP. The peritoneum is the tissue that lines the abdominal wall and covers most of the organs in the abdomen. For an overview of theories as to why PSCP is a risk for some women, including BRCA carriers, see Eisen and Weber (1998).

4. Salpingo-oophorectomy refers to removal of both the ovaries and fallopian tubes.

5. Age of onset had previously been a way to restrict the number of women for whom genetic testing was called for.

6. Several studies have suggested that because cancer is often detected at the time of prophylactic surgery, the procedure in effect functions as a screening method, thus overcoming the limits of existing screening technologies.

7. The analysis is sometimes carried out using what is called a ‘protein-truncation test’, used when clinical data are unavailable for a mutation but researchers need to know whether it is ‘deleterious’ (e.g. Risch et al., 2001).

8. As of 1968, 70% of American women with breast cancer underwent the procedure (Lerner, 2001, p. 4).

9. In 1953 Charles Huggins and Thomas L-Y Dao published a report regarding the efficacy of both oophorectomy and adrenalectomy in treating women with hormone-dependent cancer. Huggins and Dao reasoned that because hormones were detected in the urine of women after oophorectomy, removal of the adrenal glands would ensure abrupt cessation of hormone production. In order to compensate for what the researchers blithely describe as ‘an easily manageable deficiency of adrenal glands’, women in this study were treated with cortisone and desoxycorticosterone, presumably indefinitely.

10. In a survey of case studies in which doctors recommended hysterectomy in order to alleviate disease symptoms (such as those associated with endometriosis), researchers concluded that the vast majority of recommendations were based on flawed or simply non-existent knowledge regarding the patient's condition and possible alternative treatments (Broder et al., 2000).

11. Another example of recommending oophorectomy during abdominal/pelvic surgery is Seltzer et al. (1995, p. 496). According to Eisen et al. (2000), ‘There has been a lengthy debate within the gynaecologic literature regarding the practice of oophorectomy in all postmenopausal women undergoing routine hysterectomy or other abdominal surgery, with the balance of opinion favouring such surgery. A survey of members of the Royal College of Obstetricians and Gynaecologists in the United Kingdom indicated that 85% would consider the routine removal of both ovaries in a postmenopausal woman at the time of abdominal hysterectomy. Similar findings have been observed in surveys of physicians in the United States, Ireland and Italy. Proponents of this cointervention estimate that 4% to 18% of women with ovarian cancer have had a previous hysterectomy and suggest that these cases represent a missed opportunity for ovarian cancer prevention.’

12. Some researchers have suggested hysterectomy to lower the risk for cancer of the fallopian tube (fallopian tube tissue can remain behind after surgery) and to lower the risk for uterine cancer that may possibly be part of the BRCA cancer syndrome. Although some investigators have suggested that uterine cancer is part of this familial cancer ‘syndrome’ (e.g. Lavie et al., 2004), several scientists have publicly concluded that the data do not suggest a link (e.g. Karlan, 2004) and that even if there is a link, the risk is not much greater than for the general population (Kauff & Barakat, 2004).

13. According to the Massachusetts Breast Cancer Coalition, ‘[t]he fact that removal of so many healthy breasts is being hailed as “prevention” should shock us into understanding how little we really know about what actually causes breast cancer. Many women who develop breast cancer have few if any of the risk factors, and many women with risk factors never develop the disease. While for a select few women prophylactic bilateral mastectomy may be a lifesaving decision, we must keep in mind that removing the causes of breast cancer, not the breast, should remain our real goal.’

14. In a decision analysis model analyzing whether women should take HRT (Armstrong et al., 2004), researchers concluded that women should not be dissuaded from taking HRT because of increased breast cancer risk, even though they did not take into consideration the unique risks that BRCA mutation carriers face—women with BRCA mutations are more likely to develop estrogen-positive breast cancer and thus would be more susceptible to the effects of HRT (Gaber & Harman, 2004).

15. For an excellent critique of the oestrogen hypothesis as it pertains to breast cancer risk, see Krieger (1989). In a related vein, see Fausto-Sterling's (1992) work on how flawed assumptions about oestrogen have negatively influenced biological research about women.

16. Some employers have attempted to bar women from manufacturing jobs on the grounds that exposure to chemicals poses a risk to the developing foetus should a female worker become pregnant. By one estimation, as many as 20 million workers were at risk because of foetal protection policies by employers (Samuels, 1996, p. 209).

17. Further evidence of this comes from Olopade and Artioli (2004): ‘Of interest is that there have been no events in the 124 women who had oophorectomy by age 35 years, which suggests that the timing of oophorectomy may be important.’

18. For research regarding the protective effect of multiple, early pregnancies for women at risk for BRCA-related breast cancer, see Marquis et al. (1995); Holt et al. (1996); Jensen et al. (1996).

19. By ‘detect’ I do not mean detecting risk by identifying BRCA mutations in healthy women; rather, I refer to the possibility of detecting early molecular changes that signify the development of cancerous cells.

20. The theory is that organochlorines, such as the pesticide DDT and the class of chemicals known as PCBs, increase breast cancer risk because they are ‘oestrogenic’. Since oestrogen is a well-established risk factor for breast cancer, researchers have been increasingly concerned with women's body burden of organochlorine residues. In April 1993, Mary Wolff and colleagues published their finding that after adjusting for known breast cancer risk factors such as age and reproductive history, women were four times more likely to contract breast cancer if exposed to significant amounts of DDT (Wolff et al., 1993). While all study participants had some traces of DDE and PCBs (a now-banned chemical used as a coolant, lubricant, and for insulation of electrical devices) in their blood, the risk was greatest for those women showing the highest concentrations of DDT (no significant association was found for PCBs). Some subsequent studies have confirmed Wolff's results (e.g. Dewailly et al., 1994, 1997) while others have not (e.g. Krieger et al., 1994; Hunter et al., 1997).

21. Several registers of evidence suggest occupational and ambient environments are to blame: breast cancer rates vary according to geography and occupation, have been rising steadily since synthetic chemicals were introduced into agricultural and industrial production after World War II; also, women who emigrate to the US have a greater risk for breast cancer than if they lived in their countries of origin, and the women in studies such as the Long Island Breast Cancer Study Project all test positive for some amount of chemical residue in their bodies.

22. Activists, including the biologist and writer Sandra Steingraber and organizations such as the Massachusetts Breast Cancer Coalition, Ontario's Breast Cancer Prevention Project, The Silent Spring Institute, Greenpeace, the Center for Health, Environment, and Justice, and the Science and Environmental Health Network, have advocated the precautionary principle.