Skip navigation

Genewatch: BRCA genes - potential and problems

Potential and Problems

The New England Journal of Medicine had a headline back in May of which Bandolier would be proud - "BRCA genes - Bookmaking, Fortunetelling, and Medical care" [1]. It pulled together information in four papers in that issue which examined new research on these genes involved with breast and ovarian cancer. The correspondence published in September is also interesting for those contemplating how to make the best use of new genetic information.

Statistical prophecies

The discovery of the so-called breast cancer genes, BRCA 1 in 1994 and BRCA 2 in the following year, were clearly landmarks in the study of breast cancer. It was naturally hoped, indeed almost assumed, that such important discoveries would quickly be followed by rationally designed improvements in prevention and treatment of this dread disease. The message that a woman with a strong family history of breast and/or ovarian cancer who carries a germ line mutation in one of these genes carries a lifetime risk of breast cancer of about 85%, and about 60% for ovarian cancer, is a very frightening one. Not surprisingly the pressure created for the practical application of these discoveries was almost immediate and very great.

The situation was further complicated by the potential economic returns of genetic testing and the speed with which commercial companies were able to offer their services. Although the ethical and socio-economic problems raised by this were quickly appreciated, ongoing discussions between well informed representatives of the public, patient groups and the relevant professions will be required before any consensus can be reached. It is at least of some comfort to note that government regards the situation as sufficiently important to have set up (even if only after severe pressure from the professions!) both a Human Genetic Commission and a Genetics Testing Advisory Group.

Although the emerging information is of great scientific value and interest, it does have a great potential for misunderstanding and misapplication. The editorial [1] refers to the making of gene-based statistical prophecies - but not necessarily with all the evidence yet in.

Inappropriate optimism

One of the papers in the NEJM issue was a study of the prevalence of BRCA1 mutations in a population of women with breast cancer. Clinical information, family histories and blood was given by 263 women, and DNA sequencing was used to identify BRCA1 mutations. Women were consecutive referrals who were referred either because there was a strong familial risk factor for breast cancer (169 women) or breast cancer was diagnosed before 40 years of age (94 women). There was an average of 4 breast cancers per family and 1.5 cases of ovarian cancer in the women with a family history.

BRCA 1 mutations were identified in 9/124 (7%) women where there was a family history of breast cancer without ovarian cancer, 10/57 (18%) where family members had bilateral breast cancer, 28/60 (47%) with family members with both breast and ovarian cancer or a single member with both cancers.

Implication for screening

This suggests that even when subjects are selected from families with a strong history of breast cancer only about 7 in 100 will be found to have BRCA 1 mutations. So more than 9 out of 10 tests would be negative, with 10% or fewer positive, if screening for BRCA1 mutations were instituted. The implication is that BRCA1 mutations may account for fewer than half of hereditary cases of breast cancer, not much higher figures suggested previously (see Bandolier 18 ). Routine screening, even where there is a family history, will be wasteful in that it will require enormous back up, both scientific and medical, including counselling, to undertake responsibly. Furthermore, women who do not have a mutation have to be cautioned to prevent a false sense of security, because, in the striking words of the article, "negative tests are not truly negative". That is, women with negative results for BRCA1 mutation are still at risk of cancer.

BRCA1 gene

Scientists have already identified more than 200 different mutations in the two BRCA genes. But we do not know how these different mutations differ in their relative contributions to breast, ovarian and other cancers. Clearly some mutations are responsible for the very high risks identified in some families, but others are likely to be within the normal variations found in many genes (polymorphisms) and not be relevant to cancer. It is also likely that different mutations will determine different types of cancers with different pathologies and different outcomes. There are some very preliminary indications that mutations towards one end of the BRCA1 gene are more likely to lead to breast cancer while those towards the other end of the gene are more likely to give ovarian cancer. It is also suggested that mutations in the BRCA2 gene are less likely to be found in early onset breast cancer than are BRCA1 mutations.

Moreover the same mutations may behave differently in different populations of women. This emphasises the importance of modifying factors in determining the final outcome (eg, reproductive history, hormone therapy, diet, smoking and other genes which, for example, control the metabolism of hormones). We must not forget that all cancers are ultimately determined by a combination of genetic and environmental factors and the interplay of these is one of our most urgent subjects for research.

Statistical prediction

The problem of statistical prediction becomes especially worrying when irreversible decisions are based on it. The question of whether to have prophylactic mastectomies or oophorectomies for carriers of BRCA mutations is a particularly difficult one. A theoretical model [3], based on a number of statistical assumptions, estimated that prophylactic mastectomies would add 2.9 to 5.3 years of life expectancy for a 30 year old carrier of a mutation and an oophorectomy would add 0.3 to 1.7 years. Estimated gains in life expectancy decline with age, however. Gains would be minimal for a 60 year old, less than one year on average, very little more than the estimated gain in life expectancy of 8 months for a woman without a mutation.

But as correspondence suggests [4], even this reported gain in life expectancy may overstate the benefit of prophylactic surgery to the patients for several reasons:
  • The effects of surgery on quality of life are ignored.
  • Risks associated with surgery are immediate. Benefits (prevention of cancer) accrue over time.
  • The reported gain in life expectancy applies only to women who already know they have the gene mutations. Screening to find the mutation would carry its own demerits - not least those of false positive or false negative results in circumstances where prevalence may not be high.

So this information should encourage a conservative approach to prophylactic surgery. But patients are individuals rather than units of a statistical study and each needs to be given individual consideration. Where this involves changes in genes which predispose to breast cancer this means access to experts who themselves are in possession of the latest information and are able to communicate this in a digestible form to worried people who do not understand the science or statistics.


  1. B Healy. BRCA genes - Bookmaking, Fortunetelling, and Medical Care. New England Journal of Medicine 1997 336: 1449-9.
  2. FJ Couch, ML DeShano, A Blackwood et al. BRCA1 mutations in women attending clinics that evaluate the risk of breast cancer. New England Journal of Medicine 1997 336: 1409-15.
  3. D Schrag, KM Kuntz, JE Garber, JC Weeks. Decision analysis - effects of prophylactic mastectomy and oophorectomy on life expectancy among women with BRCA1 or BRCA2 mutations. New England Journal of Medicine 1997 336: 1465-71.
  4. JD Birkmeyer, HG Welch. Risk of breast cancer in carriers of BRCA gene mutations. New England Journal of Medicine 1997 337: 787.

previous or next story in this issue