Editor—Cancer genetics, and in particular breast cancer genetics, is the fastest expanding discipline within clinical genetics. Cancer referrals now constitute a third of all referrals to most clinical genetics centres. Currently there are no national guidelines on predictive testing for BRCA1and BRCA2. Several members of the same family may be seen in different centres and offered different clinical management. Such differences may in part be attributable to differences in funding of genetic services and testing at the service or research level, but it is clear that this area also involves various ethical dilemmas that may well be viewed differently by different practitioners. In order to investigate the nature and degree of variation that exists in practice and attitudes among clinical geneticists, we have undertaken a survey of all clinical geneticists in the United Kingdom who deal with cancer genetics.

Four clinical case scenarios were devised from the authors' own clinical experience to assess attitudes and practice towards breast cancer gene testing. Questionnaires were sent to 57 geneticists in the United Kingdom, representing all specialist registrar and consultants involved in cancer genetics. Each was asked to respond to questions relating to each scenario and to state the reasons for their decisions. Forty seven completed questionnaires were received (83% compliance). All clinical genetics centres in the UK were represented by at least one response. In three instances a joint response involving more than one geneticist from a centre was returned. The four clinical cases are given below. For each case the salient points raised by selected respondents for arriving at their decision are given.

Case 1. A woman has been shown to carry a pathogenic mutation in the BRCA1 gene. She is 9 weeks pregnant and requests a prenatal test to see whether the fetus also carries this mutation. Participants were asked whether they would be prepared to offer prenatal testing after appropriate counselling.

Twenty four (51%) respondents stated that they would be prepared to offer prenatal testing to the woman after counselling. Fifteen (32%) said they would not and 17% did not know what their action would be. Most of the respondents who would offer such testing indicated that if the counsellors had given all the relevant information, it was up to the woman to make a decision. Most commented that the woman's experience of cancer in the family was likely to be a strong motivating factor in the decision to request prenatal diagnosis and that counsellors were not in a position to deny this experience. Many also commented that a pregnancy could be terminated anyway for “social” reasons and that refusal to offer predictive testing would therefore not necessarily prevent a termination. The following quotations illustrate some of the opinions given.

“I would find the idea difficult but after counselling and with full information it is the couple's decision. Any other stance in a country accepting social termination is difficult.” “If we are not prepared to offer prenatal diagnosis we cannot engage in counselling.” “In some families the cancer burden is enormous. The child may not only have to grow up without a mother but also with the possibility of developing the disease herself.” “It is within the patient's rights.” “Prenatal tests are offered for other potentially treatable conditions of late onset.”

Most of the respondents who did not know what their action would be raised the issue that it depended on why she wished to have prenatal testing. They felt that testing might be appropriate if the woman would definitely proceed to termination should the mutation be detected in the fetus. If there was doubt about this, however, or the test was done for information only, then respondents would be reluctant to offer testing, since continuation of the pregnancy would be interfering with the autonomy of the child and future adult. In effect, such testing would equate to doing a predictive test on a child who might not want to know its own genetic status.

“I would attempt to dissuade her but if she was still insistent then I would feel obliged to offer a termination. There are many other disorders where I would not be entirely happy with offering a prenatal diagnosis/termination but if after careful counselling the patient still wants it then I would. After all a woman can have a termination for essentially social reasons anyway.”

Those who felt it would be inappropriate to offer prenatalBRCA testing focused on the uncertainties that still lie with such information. These are summarised by the following quotations.

“Not a lethal disorder with no treatment. If testing and no termination takes place the child could face insurance problems during life.” “Even if positive, penetrance is only 80-90% maximum. There are screening options and treatment options.” “The reasons for not doing the test are: (1) incomplete penetrance and uncertainties of penetrance in some cases, and (2) the availability of measures for early cancer detection and prophylaxis which although experimental will almost certainly improve over the next 30 years.” “Future child should have autonomy.” “Adult onset disease treatable if detected early. Prophylactic surgery a possibility.” “Effects of mutations uncertain and too far in the future.”

Some responses were more directive: “Would try to talk her out of it.”

Many respondents also felt that legally they could not offer a termination if the result showed that the fetus carried the pathogenic mutation as summarised in the following quotation.

“The 1967 Abortion Act does not apply here, therefore we have to rely on case law to do a termination for fetal abnormality. I would be concerned about the legal position here as I don't think the fetus would necessarily be classed as severely handicapped.”

Case 2. The mother of a 15 year old girl had undergone BRCA1 testing because of a very strong family history of breast and ovarian cancer and a pathogenicBRCA1 mutation had been identified. The 15 year old who was quite mature for her age and well informed about her risks was very keen to have a predictive test so that she could “plan her life accordingly”. Participants were asked whether they would offer this 15 year old girl a predictive test.

Thirty four percent (16/47) of the respondents were prepared to offer predictive testing. However, 23 (49%) were not prepared to do so. The remaining seven (15%) did not know what action they would take. Those who were prepared to offer predictive testing gave the following reasons.

“At the age of 15 a young person can be considered ‘Gillick’ competent and able to give consent to treatment or investigation. If the girl has thought it through I would be prepared to do the test.” “I believe there is benefit for her knowing at the age of 15, potential reassurance and potential lifestyle chances, usage of the contraceptive pill and planning etc if she is a gene carrier.” “After exploring her reasons for doing it now.”

For those who would not be prepared to offer predictive testing the reasons given included the following.

“Risk is not imminent. She is likely to have 10-20 years before it does become imminent and circumstances may have changed dramatically by then. It is hard to see how any immediate decisions, for example, life plans, would be influenced by knowledge of carrier status at 15.” “Thin end of wedge. Why not a mature 12 year old?” “I would advise more time for reflection. It could well affect her self image if she has a positive result and relationships with mother, and she is at a critical stage in her education which could be disrupted by a positive gene test.” “I would encourage her to wait. When life has been planned, that is, with regard to jobs, mortgages, life insurance etc.” “Her risk is negligible at this age.” “Would try very hard to dissuade her, explaining it wouldn't make any difference to her management.” “There is little information about the effects of predictive testing in young people and the psychological sequelae.” “If she thinks this information will help plan her life she needs more explanation.”

Case 3. A woman who does not fit the local criteria for genetic testing for BRCA1 orBRCA2 is insistent that she wants testing even though the limitations of the test have been explained. She asks if the test can be done privately? Respondents were asked whether they would be prepared to give her the name of a commercial company providing testing.

Seventy percent (32/47) questioned were prepared to give her this information. Most respondents felt that they could not deny knowledge on principle and it was her decision to spend her own money. Furthermore, the information is relatively easily accessible and withholding such information could heighten any anxiety. Some of the particular comments made were as follows.

“If the company find the mutation other family members have their risk altered/increased and then would probably be referred to the genetics service.” “This is already done for paternity testing. I see it as an obligation to tell the client of any other resources.”

Twenty five percent (12/47) stated they would not be prepared to give the name of a commercial company to the woman. Some felt that providing this information might be seen as an endorsement of the company and that a commercial company does not provide any form of counselling.

“I would be concerned about equity of access to the service and queue jumping by using private labs if people could afford it.” “I am prejudiced against a commercial company that does not provide general counselling support for such testing. I would point out the disadvantages of a commercial company for whose quality control operating standards and lack of counselling support I could not comment.” “We are not agents for private companies.” “If she does not fit the local criteria, the chances of finding a mutation are low. Her anxiety will not be resolved by an inappropriate and almost certainly uninformative test.”

Only two respondents (4%) gave don't know as a response and no comments were received in respect of this action.

Case 4. It was stated that a 27 year old woman attended the clinic because her identical twin sister had just developed carcinoma of the breast. Their mother also had breast cancer at the age of 45. She died at the age of 50. There was no blood or tissue available from the mother. Respondents were asked what risk would they give the consultand of developing breast cancer. They were also asked whether they would be prepared to offer any genetic testing and if so which test or tests they would offer and why?

The majority (31/47, 66%) stated that the woman's lifetime breast cancer risk was between 70 and 80%, which from published evidence seems to be the most accurate figure.1 However, some estimates were as low as 30% and others as high as 90%. Respondents were not asked how they arrived at these figures. Most of the respondents (34/47, 72%) would offer genetic testing. Their management of the case in respect of gene testing showed differences.

“Can only offer diagnostic testing, not predictive testing.” “Would do diagnostic test in sister and then predictive test in patient, because there is a chance they are not actually identical.” “Confirm they are identical and then treat as a diagnostic test.” “I would counsel the two sisters in parallel but genetically the risks are identical, the only difference is the penetrance issue of this gene in identical sibs.” “We could not give any good news, we could only give bad news.”

Only two of the respondents commented on the difficulty of interpreting a negative test.

We carried out this survey to assess variability among cancer geneticists within the United Kingdom in attitudes and practice towards breast cancer gene testing. The results of our survey clearly show that differences do exist. The differences seen in response to the clinical scenarios were, perhaps not surprisingly, most pronounced in relation to situations having the greatest ethical component.

Prenatal testing is offered for a number of genetic disorders. With the notable exception of disorders such as Huntington's disease, for which there is no disease modifying treatment, interest in prenatal testing has centred on severe diseases presenting early in life. Attitudes to prenatal testing for adult disease for which some form of treatment exists varied considerably. The issue of testing for breast cancer susceptibility genes is further complicated by the fact that such genes are not fully penetrant and that sporadic disease is common in the population.

One of the central themes of genetic counselling laid down in Peter Harper's seminal text is that counselling should be non-directive.2 3 The aim of genetic counselling is to ensure that people have the necessary facts to enable them to arrive at their own decisions. If one adopts this as the central tenet of the genetic counselling services, it is the person's decision as to whether to have prenatal testing. It may be naive to believe that all genetic counselling is truly non-directive; genetic issues are complex and counsellors' personal opinions may be apparent or inferred from the manner in which the explanation is made. We found it interesting that a large percentage of those surveyed would be reluctant to engage in prenatal counselling for breast cancer susceptibility genes and furthermore that a proportion of clinicians would actively dissuade a woman from pursuing prenatal testing if she was aBRCA1 mutation carrier. This is clear evidence against non-directive counselling.

Informed consent is clearly a central part of any predictive testing programme. The age at which a person can give consent for any medical procedure has been the subject of considerable debate in recent years. In case 2 we presented the hypothetical situation of a girl of 15 requesting a predictive test for BRCA1. Clearly a woman's risk of breast cancer before the age of 25 is small, and therefore it can be deemed that there is no immediate urgency for testing from the perspective of risk. Many of the respondents made this point, some giving it as a reason for not agreeing to offering a predictive test. However, many other reasons contribute to a person's decision to undertake genetic testing and ultimately the issue centres on at what age a person is able to make this type of decision. Some respondents felt that such testing would be illegal at this age; however, both case law (Gillick v West Norfolk and Wisbech AHA, 1986) and statute law (the Children's Act, 1989) allow children under the age of 18 to make independent decisions about themselves if they are deemed “sufficiently mature”. One respondent pointed to anecdotal evidence that presymptomatic testing experience in Huntington's disease suggests that all those under the age of 25 who were found to be carriers had major psychological problems subsequently. The only published evidence seems to point to a greater difficulty in coping with the disease when it is first learnt of during adolescence rather than adulthood.4 While deferring testing may ensure that the test is not done in haste, any directive counselling in terms of trying to dissuade her clearly runs counter to the Harper ideal of counselling.

Our next scenario concerned private laboratories. Unlike the United States, Britain has little in the way of private genetic services and both counselling and genetic laboratories are largely confined to the National Health Service. It is conceivable that things may change and there may be an increase in private genetic services. Any growth in private genetics is likely to be confined to laboratory tests and a concern here is that it may be unaccompanied by any form of counselling. Furthermore, the motivation by such enterprises is financial remuneration and hence many of those offered tests may be at a low probability of being gene carriers and at a risk not significantly different from that of the general population. It is therefore unlikely that an expansion of the activities of private laboratories will be greeted enthusiastically by clinical geneticists. Whether one should provide the address of such a laboratory if requested to do so is perhaps a different point. The responses to this question indicated a number of views relating to this issue. Some felt that giving such information could be construed as an endorsement of a private company. Alternatively, a failure to convey the address could be seen as a paternalistic attitude and erosion of free choice.

In a recent article Rosser et al 5 highlighted that there are differences in the estimation of risk made by geneticists for identical family histories of breast cancer. We also found evidence of this from the responses to case 4. While most risks were given as between 70 and 80%, estimates ranged from 30-90%. Case 4 concerned gene testing in a family with an identical twin. This case was chosen to illustrate the fact that the distinction between diagnostic and predictive testing becomes rather blurred in the context of one affected and one unaffected identical twin. Several respondents (13%, 6/47) did not comment on this and said they would proceed with predictive testing once a mutation had been detected in the identical twin. Clearly, if the twins are really identical, then only diagnostic testing is possible even in the unaffected twin. It is perhaps not surprising that there were considerable differences in participants' responses.

The field of cancer genetics is rapidly evolving and clinical practice is developing to meet the challenge of this changing field. It is perhaps not surprising that differences in clinical practices exist; however, there will be pressure for these to coalesce to similar policy. We hope that this article will draw to attention to some areas in breast cancer genetics where there are considerable differences in opinion and serve as a discussion for helping to devising guidelines for those in the field.