Twenty-five hundred years ago, the Greek city-states carried the torch of Western culture. The Greeks worshiped many gods, each with his or her temple and special authority. They believed that one temple, at Delphi, on a beautiful hillside far from Athens, stood at the precise center of the world. Delphi was dedicated to Apollo, the god of knowledge and medicine. For centuries Greek citizens, anxious about life's great questions— Will I marry happily? Will I have children? Will I live to old age?—traveled to Delphi and paid the priests to read the smoke of burning laurel leaves and divine the answers given by Apollo's grace. Early in this century, archaeologists at Delphi unearthed a stone that had lain across the entrance to the main temple. On it were carved the Greek words: "Know Thyself."
Today, genetic testing has supplanted divination, and the Delphic injunction has moved from religion to medicine. We are entering an era where we will be able to find out much more about risks of developing serious illness or of having children with genetic disorders than was even imaginable a decade ago. A decade or two from now our ability to inquire about our genetic risks will be much greater than it is today. What will we want to know? Why will we want to know it? How will the information we learn shape our view of ourselves, our spouse, or our children? Who should be allowed access to genetic information? For what purposes should genetic information be used? Is the threat of genetic discrimination so great that people should forego tests? These questions, and others like them, are the lens that will focus the emerging public debate over the uses of genetic information.
Genetic Testing: An Overview
Genetic testing began in the early 1960s when we started to screen all newborns for phenylketonuria (PKU), a rare genetic disorder (1: 12,000
births) that causes mental retardation. By promptly placing affected infants on a special diet which is low in phenylalanine, it is possible to avert otherwise certain retardation. Today, virtually every newborn in the western world is tested for PKU. The program has been so successful that few doctors have ever seen a person who is mentally retarded because of PKU. During the late 1980s I was the medical director of a large residential institution for mentally retarded persons in which lived several adults with untreated PKU. When I arranged a visit by a group of Boston physicians who cared for children with PKU (all of whom had been diagnosed at birth and successfully treated) to observe my patients, the doctors were astounded by the severity of the mental retardation. The persons with untreated PKU, who had no use of language and suffered from seizures, had the tiniest genetic error—just one or two misspelled DNA letters out of billions—but an error with devastating consequences.
Newborn genetic screening, which today has expanded to test millions of children for (depending on the state) three to ten severe but treatable metabolic disorders, has posed relatively little controversy. This is in large part because the tests are highly accurate, and there is much we can do to avert disaster in the children who are born with these rare disorders. However, newborn screening, once limited to a few metabolic diseases, has entered an era that will be marked by dramatic expansion of the kinds and numbers of tests that are run. Currently, the expansion is being driven by the use of powerful, inexpensive technology called tandem mass spec-trometry.
Tandem mass spectrometry (TMS) analyzes a tiny volume of blood eluted from a spot on a special kind of paper. The mass spectrometer assesses the size of molecules in the blood, creating a kind of chemical fingerprint. When the machine reports an unusual molecule, the technician (or more likely built-in software) can compare it to a database of metabolites associated with a variety of rare genetic disorders and infer a diagnosis. In 1999 Massachusetts became the first state to include TMS as a routine part of its mandatory newborn screening program. An advisory committee to the Commissioner of Public Health, on which I served, concluded that a disorder known as medium-chain acyl CoA deficiency was sufficiently common and sufficiently treatable to warrant screening. In addition, Massachusetts launched a pilot, voluntary screening program that uses TMS to look for chemical signals indicating that an infant may be affected with one of an additional 20 disorders. About 99% of new mothers who are invited to participate in the pilot program do so.
TMS has arrived at an excellent moment, both because of the benefits it offers and because it is a surrogate for a not-too-distant time when all newborns will routinely undergo DNA screening that may ascertain and compile thousands of questions about their genetic health. A limited DNA testing program is already part of newborn screening in Massachusetts; it is used to confirm the diagnosis of cystic fibrosis in infants with a positive test on a simple biochemical screen. In debating the proper uses of TMS, we will rehearse the future use of DNA testing of newborns. Today, DNA testing is too expensive to use on all newborns, but costs are decreasing rapidly. With the advent of a low-cost DNA array chip (a device on which are anchored thousands of short segments of DNA each designed to look for a particular mutation), we will have to decide what kinds of information it is appropriate for the state to find out about people.
No one has ever challenged the constitutionality of laws that require babies to be tested for these rare disorders. However, if newborn genetic testing expands to identify infants with less severe problems (which it is almost certain to do), it is inevitable that some parents will argue that compelling them to submit their children to genetic testing is an invasion of privacy. And so it is, but the real question is whether a mandatory test confers a benefit to society that outweighs the potential harm to individuals. Historically, the courts have accorded the state significant latitude in the exercise of the "police power" on behalf of the public health. One could argue, however, that programs to identify children at risk for genetic disorders do not seek to find persons who pose a risk to the public health, and there is thus an insufficient basis to compel testing. In addition, one could argue that any test that merely produces information about a degree of risk for an adult-onset disorder or for a non-life-threatening condition falls far short of justifying mandatory testing.
Prenatal testing raises even tougher ethical issues than does genetic screening of newborns. In our society, pregnant women are generally offered two tests aimed at identifying fetuses with birth defects. Those women who will be 35 or over at delivery are advised about the use of amniocentesis to find out whether the fetus has a chromosomal disorder, mainly Down syndrome. Virtually all pregnant women are offered a blood test to determine whether the fetus might have spina bifida (improper closure of the spinal column). More than a third of pregnant women over 35 undergo amniocentesis (aspiration of amniotic fluid to obtain fetal cells), and about 80% of those who learn they are carrying a fetus with Down syndrome decide to end the pregnancy. About three-quarters of pregnant women take the blood test (which measures proteins made by the fetus that cross the placenta into the mother's blood) to screen for spina bifida. About three-quarters of those women who learn that the fetus has open spina bifida (thus making it likely that they will have moderately severe birth defects) terminate the pregnancy. These are painful decisions, not the least because doctors usually cannot predict the severity of the impairment that the fetus is likely to have.
The third major type of genetic screening in use today seeks to identify persons who carry genes for recessive disorders. Should these otherwise healthy persons marry another carrier of a mutation for the same disorder, the couple face a 1 in 4 risk in each pregnancy of having a child with the disease. In the United States, population-based carrier screening has focused on identifying persons at risk for bearing children with sickle cell anemia (mostly African-Americans and people of Mediterranean origin) and persons at risk for bearing children with Tay-Sachs disease (mostly Ashkenazi Jews). Both the carrier screening programs and the outcomes have been dramatically different. Despite more than 20 years of screening, there has not been a reduction in the expected number of births of children with sickle cell anemia. During the same period, the births of children with Tay-Sachs disease in the Jewish community has fallen by about 95%. There are many reasons for this disparity, including differences in access to health care, and cultural differences about the use of selective abortion, but the major one is the difference in the severity of the two diseases. Tay-Sachs disease is a uniformly fatal disorder of early childhood. Persons with sickle cell anemia have many medical problems, but a child born with the disorder today has a good chance to live a productive life into his or her 40s or beyond.
Prenatal testing and carrier screening have made it possible for women at risk to avoid the birth of children with two of the most common birth defects and two of the more common recessive disorders. For example, in England, prenatal screening and selective abortion have led to a 95% reduction in the birth of children with spina bifida. Many women now regard genetic testing as an important component of their health care. From a public health perspective, the picture is more complex. One could argue that western society has embarked on a course that seeks to avoid the births of persons with certain kinds of disabilities. Such a decision may seem fairly straightforward in cases of severe, incurable disorders (e.g., Tay-Sachs disease), but consider a future in which we will examine hundreds of genes to find out about a huge variety of risks. Will some women abort a fetus because it is at high risk in mid-life for ovarian cancer? Will others want to abort if they learn that the future child will almost certainly be severely obese? Should tests that merely predict health issues in adulthood even be offered as part of carrier screening or prenatal diagnosis?
As DNA-based predictive tests proliferate, some women and couples will use them in ways that others will find offensive. In the United States, each year about 1,000,000 women terminate pregnancies, in more than 98% of the cases because they do not wish to bear a presumably healthy child. Public opinion, which remains bitterly and intractably divided over abortion, will surely be inflamed over the growing practice of selective abortion—the termination of a pregnancy based on a particular prenatal diagnosis. But does not the right to reproductive privacy include a woman's decision to terminate a pregnancy even if it is for reasons that many persons abhor?
For 30 years genetic testing, whether it be newborn screening, prenatal diagnosis, or carrier screening, has focused on identifying risks for serious disorders that would be present at birth. The revolution in molecular genetics will change the nature of the data that could be made available about us. For example, many of the DNA-based tests that could be used in the future screening programs will not diagnose a disorder. Rather they will recalculate the odds of developing a disorder. In so doing, they will raise new and difficult questions about what we might want to know and what use we can make of such knowledge. The following vignettes illustrate some of the dilemmas we will face.
During the last 10 years, Peter Loomie, a soft-spoken computer programmer, watched his father and one of his uncles die in their 50s of a rare form of Alzheimer disease. In his family, the disease is due to a mutation in a gene on chromosome 14. Peter, who is in his mid-30s, is married but has no children. A DNA test for the mutation is available, and he has spent many sleepless nights wondering whether to take it. He calls the test the most important coin toss in his life. There is a 50% chance that he inherited the mutation from his father, and a 50% chance that he did not.
The test results might redefine his life. If he takes the test and it shows that he inherited the mutation, Peter does not think he would have biological children because he cannot bear the thought that he might pass the mutation to them (a 1 in 2 risk). Because he would likely become ill a decade hence, he doubts whether he would be willing to adopt. Although he is happily married, Peter secretly fears that if he tested positive, his wife might leave him. He admits that he has no basis for this fear. His wife knew about the family illness before she married him, and since the predictive test became available, she has steadfastly maintained that the decision about whether or not to be tested was his alone.
If Peter takes the test and it reveals that he did inherit the mutation, Peter says he would quickly load up on insurance. He and his wife will stop thinking about their retirement years and redouble their efforts to enjoy the present. If the test shows that he did not inherit the mutation, Peter's risk for early-onset Alzheimer disease vanishes (his risk for the much more common form of Alzheimer disease that strikes in old age is still the same as that faced by the rest of the population). The fear that he and his wife struggled with concerning childbearing—a fear of passing on a "bad" gene to kids—would also evaporate. The way they think about jobs, savings ac counts, travel, insurance, and myriad other matters would dramatically change.
Peter is at risk for an exceedingly rare disease, but there is also a predictive test for the common form of Alzheimer disease, the type that affects up to one-third of Americans who live into their 80s. About 2% of us—more than five million people—carry two copies of a gene variant called apoE4 which confers a much higher than average risk of Alzheimer disease (see Chapter 19). Those of us who are born with two copies are not certain to develop Alzheimer disease, but are much more likely than average to do so and at a somewhat younger age (mid-60s). The test for apoE4 is inexpensive and accurate. Should doctors use it as a susceptibility test? Currently, they do not. It is recommended for use only as a confirmatory test in people who are already showing signs of possible Alzheimer disease.
How should doctors respond to patients like Martin Dulbecco, a 45-year-old banker whose father developed Alzheimer disease at 64 (and who had two copies of apoE4), and who insists on being tested? Should he be provided with this information? Martin told me that since he already lives each day believing that he shares his father's fate, the test can only help his situation. If he does not have two copies of apoE4 he will conclude that he will not develop Alzheimer disease in his 60s; if he does have two copies, it will merely confirm what he "knows." How tightly should physicians exert controls on access to genetic information? Do physicians have any right at all to deny access to such tests?
The first highly informative DNA-based predictive test—for Huntington disease (HD), a late-onset, untreatable brain disorder that slowly kills—became available in the mid-1980s. Although he had not taken the test, for years Richard Coe, a 37-year-old tennis pro, whose father died of HD, lived as though he had the HD gene. He lived the motto, "Eat, drink and be merry, for tomorrow we die." As Richard tells it, he spent freely and was always deeply in debt. He had a vasectomy when he was 21, had lots of girlfriends, but never a serious relationship, took up hang gliding, and parachuted regularly. He lived on the edge and loved it. Then, he "made the mistake of falling in love." After two years of internal debate, he took the HD test. To his shock, he learned that he had not inherited the mutation. "My life was ruined," he recalls laughingly. "I had to get married, re verse the vasectomy, stop hang gliding, pay off the credit cards, and start being polite to my boss." Richard's story shows that there are vast differences in how individuals react to genetic risk and to learning whether that risk could manifest.
Genetic testing created a dramatically different problem for Ada Boyd, a pediatrician who was caring for a little girl with cystic fibrosis. As part of her interest in how different mutations within the CF gene shape the severity of the disease, Dr. Boyd tested the DNA of the girl's parents. The test showed that the husband did not have either of the mutations that were present in the little girl's DNA. The inference was inescapable. The husband is not the father of the child. To whom (if anyone) should Dr. Boyd disclose this information? Because the couple sought her help together, she believes that she owes an ethical obligation to each person. If she discloses the finding of nonpaternity to the husband, she may threaten the marriage.
Experienced clinical geneticists confide that the inadvertent discovery of nonpaternity is an uncommon, but not a rare, event in their practices. Although there are few good data, population geneticists estimate that un-revealed nonpaternity is present in about 1-5% of families. How is such a finding handled in the genetics clinic? Surveys indicate that more than 90% of clinical geneticists who discover nonpaternity inform the woman, but not the man. These surveys, however, were conducted before the public's now steady exposure to genetics. Today, if the husband were to ask about the implications of his not having a mutation, it would be nearly impossible to avoid telling the truth. Is there a valid ethical difference between not offering the truth and lying? To me the current practice seems difficult to defend and will almost certainly have to be revised as consumers become more sophisticated about genetic testing.
Discovery of nonpaternity can also confound genetic research. I was consulted by a scientist who had been studying a large family burdened with a rare genetic form of cancer. In mapping the culprit gene, he had discovered that one of several adult children could not possibly have inherited the risk for cancer because she was not the daughter of the parent who had passed on the mutation. Telling her that she could not be at risk would lift a great burden from her shoulders, but it could add a new one. If she asked why, the scientist would have either to lie or to reveal that the woman is not the daughter of the man that she has presumed for 40 years is her father. After much soul-searching, the scientists and doctors decided to hide the discovery of nonpaternity. They even sought and obtained the approval of the editor of the journal in which they published their research to alter some genetic facts about the family so that no one could draw the correct inference.
One of the toughest ethical dilemmas that I have faced in clinical genetics began accidentally. One day in 1986, while examining a mentally retarded man with a high fever, I noticed that he had a prominent forehead, simple cupped ears, and large testicles—features associated with a now commonly recognized disorder known as Fragile X syndrome (named for the fact that, depending on how the cells are prepared for study, the X chromosome may look broken under the microscope). Without thinking through the implications of my decision, I ordered a special chromosome test. Two weeks later, I learned that the test confirmed that this 47-year-old man did have Fragile X syndrome, an X-linked genetic disorder. Knowing that this knowledge had important implications for the man's two nieces, both of whom were recently married, I called his sister (their mother) who was his guardian. I told her the diagnosis, and asked permission to talk to her daughters.
The woman was outraged. Without explanation, she forbade me to contact her daughters and slammed down the phone. Over the next month, I tried repeatedly to meet with her, even sending a registered letter. Finally, a social worker who had the sister's confidence figured out what was going on. The woman, who had for 40 years believed that her brother was retarded due to a fall from a crib, had suddenly and dramatically had the view of her brother's illness reshaped. She now had to confront the fact that her brother was retarded due to a gene that he had inherited from their mother. She may or may not have realized that there was a 50-50 chance that she too had inherited the gene (which tends to affect women much less severely) and, that if she had, there was a 50-50 chance that her daughters had in turn received it from her. If so, in each of her daughter's pregnancies they would have a 1 in 4 risk of bearing a son with Fragile X syndrome.
Why would the woman not talk to me? Why had she told the social worker that she would never tell her daughters about Fragile X syndrome? Her reason was crystal clear. Her daughters, she knew, were pro-choice, and would seek prenatal diagnosis and would abort an affected fetus. She, however, was vehemently pro-life. As she said to the social worker, "I will never be party to telling my daughters something that could lead to the abortion of a grandchild!"
If I had not ordered the blood test to make the diagnosis, I would never have confirmed my suspicions and upset the sister's world. But, now I did know, and, given her adamant position, I was the only person who could alert the two young women to their risks, before they became pregnant. Yet, if I did tell them, I would be disobeying the clear instructions of my patient's guardian and violating his right to privacy. The man's sister did not relent and I did not warn her daughters. I hope that before they became pregnant the young women told their physicians about their uncle, and that the doctors pursued that family history.
I do worry that someday one of my former patient's nieces will tell me she has a son with Fragile X syndrome, and ask me why I never warned her of her risk. Could she sue me? Faced with a strict order not to breach confidentiality, there is little chance that I could be successfully sued for failure to warn her about Fragile X. True, obstetricians who fail to warn pregnant women over 35 of their age-associated risk for bearing a child with Down syndrome are regularly sued for "wrongful birth" (see Chapter 8) when such a child is born, but there are two key distinctions. Neither of the nieces ever sought me out for care and, even if they had, I would clearly be under an obligation not to reveal confidential facts about my patient without his (or his guardian's) permission.
As it spreads through medicine, genetic testing will create many dilemmas about sharing sensitive information within families. At first, most people are surprised that this could be a significant problem, but every family, especially every extended family, has its own peculiar dynamics, its share of skeletons in the closet, and its share of relatives who don't get along. Patients will sometimes tell doctors not to disclose facts that the physicians think should be shared. Perhaps to deal with this issue, we will someday have to amend our notion of confidentiality concerning the transmission of genetic information within families. Our ancient principle of patient-centered confidentiality may evolve into a family-centered principle—one that permits physicians to make limited disclosures to a few relatives if the medical issue is significant. This is a subjective standard, one that in practice will vary from doctor to doctor, but I know of no better approach.
Those who adhere to an absolute principle of confidentiality, a position I disagree with but respect, may find my position more acceptable if it is coupled to another. Before they test patients, doctors must tell them that should the test uncover genetic facts of great potential importance to others, their routine practice is to make sure a warning is made. The patient who is offended by this position may then seek testing through a different physician who does not feel so compelled. The only international survey (conducted by my colleague, Dorothy C. Wertz) of clinical geneticists on this issue indicates that many are beginning to rethink confidentiality. Among 2000 geneticists in 37 countries, Wertz found that the majority would tell an individual's relatives about a genetic diagnosis, over the objection of the patient, if the relatives asked. In the United States, clinical geneticists split down the middle on this issue; few geneticists report they would actually seek out and warn the relatives. About half say that they would disclose if they were directly asked.
Should physicians have the right to refuse to test at-risk individuals who want to be tested? To suggest that a physician could ethically refuse to perform an available genetic test may seem paternalistic, but medicine is replete with instances of physicians denying patients' requests. To cite the most common example, most physicians will not prescribe antibiotics to people with colds that appear to be caused by viruses, despite the fact that many such patients visit their doctors expecting to get such a prescription.
Consider presymptomatic testing for Huntington disease. For a while in the mid to late 1980s, there were less than 20 academically based laboratories where one could be tested. These labs made it their policy not to test at-risk persons who were under the age of 18 and not to test an identical twin unless the co-twin also wanted to know whether he or she had inherited the mutation. The geneticists reasoned that, because knowing whether one had inherited the allele carried with it absolutely no medical benefit, it was better to let each at-risk child grow up and decide on his or her own whether to be tested, rather than have a parent do it. In the case of identical twins, since testing one person diagnoses two, the early view, to which few doctors adhere today, was to test only with the permission of both.
Unless there is clinical benefit associated with the knowledge, geneticists usually balk if parents request that a child be tested because the parents want to know. One especially challenging case I was called about involved a young woman who had learned that she carried a gene that put her at high risk for breast or ovarian cancer. She wanted to test her two daughters, who were six and eight years old. Why? She wanted to use the results to decide whether to have more children. If either child had inherited the normal allele the mother wanted no further pregnancies. If both children had inherited the cancer gene, she wanted to become pregnant again, hoping to have a child without this added risk. Reasoning that the children would not be tested for their own benefit, doctors refused to comply with her request. They were in part influenced by the suspicion that if the woman learned that both daughters were carriers, she might seek prenatal testing and abort a female fetus who had the carrier gene.
Although the vast majority of abortions occur because women do not wish to be pregnant, in the future, a steadily growing fraction will be as a consequence of a genetic diagnosis. What constitutes a sufficiently severe genetic risk to justify prenatal diagnosis and selective abortion? Should the choice belong absolutely to the pregnant woman? The classic debate that focuses on this topic is sex selection. For many years in the United States, clinical geneticists (except in cases of sex-linked disease) refused to perform amniocentesis to determine fetal sex for women who planned to abort if the fetus was not of the desired gender. Sex selection (based on fetal ultrasound and abortion) is widely practiced in some societies, especially in parts of India. Even in the United States, some physicians, driven by respect for patient autonomy, no longer feel that they have the right to refuse genetic testing to determine fetal sex.
Physicians, geneticists, and genetic counselors are acutely aware that they have been given the keys to a library full of powerful information.
They do not want to sit on Solomon's throne. What should they do? As genetic tests become ever more widespread, physicians will seek public guidance on the proper uses of the test results, as is already happening in regard to regulating the use of genetic information by insurers and employers. In the absence of regulations, physicians will be ever more likely to accede to the requests put to them.
8-cell human embryo. (Photo courtesy of Sherman J. Silber, M.D., Director, Infertility Center of St. Louis.)
Was this article helpful?