Genomic Messages: How the Evolving Science of Genetics Affects Our Health, Families, and Future - George Annas, Sherman Elias (2015)
Chapter 6. Genomic Messages from Fetuses
If disclosure and consent are ever to have meaning,
physicians must learn to manage uncertainty better.
—Jay Katz, The Silent World of Doctor and Patient (1984)
The increasing ability of science to prenatally diagnose genetic disorders and congenital anomalies in the fetus is the subject of this chapter. Progress has proceeded so fast in the past decade that there is little doubt the technologies we summarize will have evolved further by the time this book is published. Nonetheless, the social, legal, ethical, and public policy issues raised by increasing the number and types of genetically determined or genetically influenced traits that can be detected in the fetus will not change and are ultimately more important than the science in determining what you can and should do with this newly acquirable genomic information. The opening quotation from Jay Katz, the world’s leading authority on informed consent, is meant to suggest that even with all our new technologies, uncertainty is still at the heart of decision making during pregnancy.
As you read this chapter, envision yourself as pregnant and ask yourself the following questions: What are the “boundaries,” if any, that should restrict or prohibit prenatal diagnosis (for example, the severity of the condition for which testing is performed, stage of pregnancy, likely planned actions if an abnormality is detected, use to determine fetal sex or other nondisease characteristics, such as eye color, height, athletic ability, musical talent, or intelligence)? When you see your physician and he or she offers a prenatal genetic test, do you want to be told simply that your physician “recommends” the test, or would you rather have your physician explain the benefits and risks of the test in a “neutral” (unbiased) manner and leave the decision of whether to have the test strictly up to you?
Other questions you should be thinking about involve the level of detail. How much information do you need to make an “informed decision” about prenatal genetic diagnosis? Do you want to know about all the symptoms, treatments (if any), underlying causes, and prognosis of each disorder? Would you want to have this information before the test is performed, or would you rather wait until the test results are back, dealing with these issues only if they are relevant to you or your family? Do you want to know the likelihood that the test will show a problem in the fetus and the accuracy of the test? If you are told that the fetus has a serious genetic abnormality or anomaly, what would be the best course of action for you to take (that is, continuing or terminating the pregnancy), and who should make this decision? What additional information would you want from your physician? Would you go to the Internet before deciding? Do you think the more information you have the better? These are all difficult questions, and your answers will depend not only on medical factors but also on your personal beliefs and values.
In 2013 the American College of Medical Genetics and Genomics (ACMG) opposed legal restrictions on abortion following prenatal diagnosis. The college noted that the entire practice of medical genetics is providing patients with information to enable the pregnant couple “to choose a safe and personally acceptable management plan,” and concluded that “termination of pregnancy for genetic disorders or congenital anomalies that may be diagnosed prenatally is a critically important option.” This statement was prompted by the passage of a North Dakota law outlawing abortion for a “genetic abnormality,” defined as “any defect, disease or disorder that is inherited genetically.” The law also has its own nonexclusive list of prohibited indications for abortion: “Any physical disfigurement, scoliosis, dwarfism, Down syndrome, albinism, amelia, or any other type of physical or mental disability, abnormality or disease.” This law and proposals like it are a direct response to the increasing number of genetic tests that can be performed on fetuses—and thus the increasing number of conditions that might lead pregnant women to terminate their pregnancies. North Dakota governor Jack Dalrymple said he signed this law to challenge the “boundaries of Roe v. Wade,” the 1973 U.S. Supreme Court opinion on abortion.
Roe provides that a pregnant woman, in consultation with her physician, has a constitutional right to terminate her pregnancy prior to fetal viability, and that she retains that right even after fetal viability if her “life or health” is endangered by continuing the pregnancy. The state may impose some requirements on previability abortions, such as requiring informed consent and a twenty-four-hour waiting period, but only if these conditions do not “unduly burden” women by actually preventing them from obtaining abortions. The chances that even the current Supreme Court would find the North Dakota law constitutional approach zero because the law is so extreme. For example, it would prohibit abortions even in cases where the fetus is nonviable because of anencephaly (absence of a major part of the brain) and cases where the baby would die a slow, painful, and inevitable death, such as Tay-Sachs disease. Nonetheless, just as society is unsympathetic to abortion when the fetus is the “wrong” sex, we believe society is likely to continue to support “genetic abortion” only if it is limited to serious genetic conditions. Down syndrome has historically been considered the classic serious condition, and as a descriptive matter but not an ethical one, new conditions are often looked at as being more or less serious than Down syndrome. As for Down syndrome itself, 60–90 percent (the number varies by survey and depends on a number of factors) of women whose fetus is diagnosed with Down syndrome elect to terminate the pregnancy.
Society has not had to decide what genetic conditions justify abortion, but as we will discuss, the rapid expansion of genetic tests of the fetus makes this a compelling contemporary quandary. Can we simultaneously reject eugenics and discrimination, while providing pregnant women with ever more genomic information about their fetuses? Before we can begin to answer this question, it is critical to understand how we got to this point and what values underlie the practice of prenatal diagnosis.
Prenatal genetic diagnosis began in the mid-1960s with the development of amniocentesis, a procedure in which a small sample of amniotic fluid surrounding the fetus is withdrawn from the uterus via a needle inserted through the pregnant woman’s abdomen. Either the fluid or the cells suspended in it are used to detect serious fetal genetic abnormalities. Among the first to perform amniocentesis in clinical practice was Albert Gerbie at Northwestern University, Sherman’s early mentor and the obstetrician who delivered Sherman’s two sons. In the beginning, amniocentesis was considered far too dangerous because it was thought to increase the risk of miscarriage and serious complications in the pregnant woman.
Gerbie and others who pioneered prenatal genetic diagnosis using amniocentesis did so at considerable peril to their own careers. Research on amniocentesis occurred before Roe v. Wade, at a time when in most states women had great difficulty in obtaining authorization for an abortion, even when a serious fetal abnormality was diagnosed. Twenty years after Roe, a committee of the Institute of Medicine (IOM) suggested clinical guidelines for prenatal diagnosis. The committee prophetically anticipated that many more genetic tests would be developed, and “eventually technologies will be available to simultaneously test for hundreds of different disease-causing mutations, either in the same or different genes.” The committee’s recommendations, which remain valid today, included these:
• Patients must be fully informed about the risks and benefits of testing procedures, their possible outcomes, and alternatives.
• Prenatal diagnosis should only be offered for the diagnosis of genetic disorders and birth defects, not for minor conditions or characteristics, or for fetal sex selection.
• Education before and after prenatal screening should be available to patients, and ongoing counseling should be available following pregnancy termination.
• Reproductive genetic services should not be used for the eugenic goal of “improving” the human species.
The recommendations became the foundation upon which professional organizations, most importantly the American College of Obstetricians and Gynecologists (ACOG) and the American College of Medical Genetics and Genomics (ACMG), built their own guidelines for prenatal diagnosis. The amount of genomic information that can be obtained from fetuses, as well as the variability and uncertainty of genetic findings, is so great that it is often difficult for anyone to interpret their meaning. Prenatal diagnosis can, however, sometimes identify a serious condition, and this information could lead a couple to terminate a pregnancy. The ability to determine the health of the fetus also permits couples that would not otherwise take a chance on having a baby because of the high risk of a serious genetic condition, such as Tay-Sachs disease, to have children without risking the birth of an affected child.
Genomic information about fetuses can also include information that is unclear and even contradictory. Such information can increase uncertainty and result in confusion, anxiety, and difficult decision making for everyone—physicians, patients, and families alike. In a debate on the future of the new genetics with Francis Collins at a national genetics meeting in New Orleans in 1993, George (who was asked to address the “dark side” of the Human Genome Project) suggested that the proliferation of genetic and genomic information about fetuses would, at least temporarily, lead to a marked increase in abortions. Geneticist Paul Billings had made this suggestion a year before, saying, “For now and for the foreseeable future, a major benefit derived from genetic information by families and individuals is the possibility to prevent the birth of other gene carriers by utilizing selection abortion.” At the same congressional hearing in which Billings made this suggestion, the then head of the National Institutes of Health (NIH), Bernadine Healy, said she did not think the genome project was directed at abortion. Collins, who is currently the head of NIH, would undoubtedly take the same public position. Even in 1993, Collins was not happy to hear this suggestion, and both he and Healy were certainly correct in suggesting that increasing the number of abortions was not the goal of the Human Genome Project. Nonetheless, for the foreseeable future the new genetics will be used primarily to enable much more detailed prenatal screening and diagnosis, and will therefore increase the number of abortions. Preventing these abortions is the stated goal of laws like the one passed by the North Dakota legislature.
Most commentators and many physicians use the terms screening and diagnostic testing interchangeably. Screening usually refers to looking for problems or conditions in the general (healthy and asymptomatic) population, whereas a diagnostic test or testing is used to see if someone suspected of having a disease or condition actually has it. Examples of screening the fetus during pregnancy include a combination of maternal serum biochemical markers and ultrasound findings to determine if a fetus is at increased risk of having Down syndrome. A diagnostic test, on the other hand, identifies individuals or fetuses that actually have a specific disease or condition; it diagnoses a condition. There are approximately 4 million births annually in the United States. All these pregnancies are candidates for prenatal screening and diagnosis. Our best estimate is that about 350,000 pregnant women undergo prenatal diagnosis annually in the United States, and about 70 percent of all pregnant women undergo screening for fetal Down syndrome. It is worth stressing that all these tests, both screening and diagnostic, are voluntary and should never be performed without the informed consent of the pregnant woman.
Vanessa and Steve (not their real names) were treated by Sherman. We tell their story because it illustrates an inherent paradox in modern obstetrics: in trying to make pregnancy and childbirth safer for women and increase the probability of having a healthy baby, it is necessary to convey information (and perform tests) that will likely significantly increase the anxiety of prospective parents. Some physicians tend to go overboard and give an exhaustive list of all possible problems just in case lightning strikes; this is often referred to as “defensive medicine.” The thinking is that it is better to “overinform” just in case something unanticipated goes wrong, so the physician cannot be blamed for failing to mention the possibility. Of course, simply increasing the amount of information conveyed risks confusing the patient, losing perspective of the most important and common issues that should be considered, and inflicting needless anxiety. How much information someone really needs to be an informed patient is not an easy question. Each patient really is different, and there is no “right” answer that applies to all patients. Informed consent has basic minimums but ultimately should also be personalized. The minimum is that all patients must be provided with information that might lead a “reasonable patient” to reject the screening, but information should be tailored to the individual’s risks, including the health of prior children and the ethnicity of the couple.
As you read the story of Vanessa and Steve, ask yourself if you think Sherman struck the right balance between giving too much or too little information. Would you have wanted more or less information than Sherman provided? Would Sherman’s use of more simplified language have been condescending, or was his expectation of the couple’s medical knowledge too great?
Vanessa, a twenty-nine-year-old social worker, and her husband Steve, a thirty-two-year-old electrical engineer, were very excited; they had just learned that they were going to have a baby, their first. When Vanessa was about ten weeks pregnant, Sherman saw her and Steve at her first office visit. Sherman began with a general discussion of the dangers of cigarette smoking, alcohol and drug use, and excessive weight gain during pregnancy, as well as the importance of prenatal vitamins and proper diet, and he recommended an influenza vaccination. He told the couple that the laboratory tests he routinely obtains on all prenatal patients include a Pap smear, blood type, Rh factor and antibody screen, a one-hour glucose challenge test for diabetes, urine culture and protein assessment, and screening for infections, including syphilis, gonorrhea, chlamydia, hepatitis B, Group B strep, and HIV. He also told them that most women have several ultrasounds during pregnancy, and that this was useful in determining the gestational age, the number of fetuses, fetal viability, the location of the placenta, fetal growth disturbances, and amniotic fluid volume, as well as detecting major fetal anomalies.
Vanessa and Steve completed a detailed intake questionnaire that included a family history of genetic disorders and birth defects. Sherman explained that the American College of Obstetricians and Gynecologists recommends that invasive prenatal testing—first-trimester CVS or second-trimester amniocentesis—for chromosomal abnormalities be made available to all women, regardless of age (the older guideline was to test at age thirty-five). He explained that amniocentesis involves inserting a needle into the uterus and amnion (the sac surrounding the fetus) under ultrasound guidance and withdrawing a small amount of amniotic fluid. Chorionic villus sampling (CVS), which is also performed under direct ultrasound visualization, entails collecting a small sample of placental tissue either by inserting a thin flexible tube through the vagina and cervix into the uterus or by inserting a needle through the woman’s abdomen. Both CVS and amniocentesis carry small risks of complications (less than 0.5 percent), most importantly miscarriage, infection, leakage of amniotic fluid, or vaginal bleeding after the procedure. Fetal injuries are very rare.
Sherman went on to say that various noninvasive first- and second-trimester fetal screening tests (not diagnostic) can be used rather than going directly to invasive prenatal testing. Specifically, to achieve a higher rate of Down syndrome detection (and a few other chromosome abnormalities) and reduce the need for invasive tests, an “integrated screening” approach can be used. This approach combines first-trimester ultrasound measurement of a fluid-filled space behind the fetal neck, called the nuchal translucency, with maternal blood biochemical markers. These first-trimester screening tests are then followed by additional maternal blood markers, collectively called a “Quad test,” performed in the second trimester. These tests can detect about 95 percent of all fetuses with Down syndrome.
An alternative to integrated screening would be for Vanessa and Steve to choose between two “sequential screening” strategies. Sherman said that his own preference is the “stepwise sequential strategy.” If the risk for Down syndrome was greater than 1 in 30, based on a first-trimester serum screening and nuchal translucency measurements, Vanessa would have a second-trimester Quad test, after which the results of both tests would be combined. She and Steve could then decide (based on the risk of Down syndrome) if they wanted an amniocentesis. The second approach is “contingent sequential screening.” Vanessa would be assigned to one of three Down syndrome risk groups based on the first-trimester screen results. If the risk was greater than 1 in 30, CVS or amniocentesis would be performed; if the risk for Down syndrome was “intermediate” (between 1 in 30 and 1 in 1,500), second-trimester screening would be performed. If the risk was lower than 1 in 1,500, no further testing would be done. This approach lays out a plan at the outset and lets the couple avoid having to make a series of decisions as the pregnancy proceeds.
Sherman also said that during the second trimester a blood test, called maternal serum measurement alpha-fetoprotein, or MSAFP, is routinely offered as a screening test for certain neural tube defects. The most common neural tube defects are spina bifida (in which the spine is partially open, which can lead to nerve damage and varying degrees of disability) and anencephaly (in which a large part of the brain and skull fail to develop, and which is almost always fatal within the first few days of life). If the MSAFP screening test is positive, targeted ultrasound imaging of the fetus, and possibly amniocentesis, would be recommended for further evaluation. If they decided to have an amniocentesis, MSAFP screening would not be necessary because alpha-fetoprotein testing would be done on the amniotic fluid sample.
Finally, Sherman told the couple about a newer test called “noninvasive prenatal screening” (NIPS), which is based on analyzing fetal DNA found in the mother’s blood. The test is highly accurate (greater than 99 percent) for detecting Down syndrome and a few other chromosome abnormalities, but it would not provide information about any other fetal disorders. Although highly accurate, NIPS is still considered a screening test, so if the results indicated a fetal problem, a confirmatory CVS or amniocentesis would be recommended. Sherman told the couple that the interpretation of NIPS was less complicated than maternal serum screening and would be less likely to indicate the need for a CVS or amniocentesis. He therefore suggested that the couple consider NIPS as their first choice. The first-trimester ultrasound and a second-trimester MSAFP blood test would still be recommended.
Sherman concluded this part of the discussion by telling Vanessa and Steve that a couple’s decision whether to undergo fetal screening and what tests to have is based on many factors, including their tolerance for uncertainty about the health of the fetus, the risk of pregnancy loss from an invasive procedure, and the consequences of having an affected child. He told them that if they wanted, they could go directly to having either a CVS or amniocentesis as long as they understood and accepted the risks.
Based on their ethnicity, Sherman also recommended additional screening for the couple. Because Vanessa was of Greek descent, he suggested a complete blood count and hemoglobin electrophoresis to determine if she was a carrier for beta-thalassemia, a severe form of anemia (Steve would also be tested for this if she was positive). Steve was of Ashkenazi Jewish descent; because of his ancestry, Sherman recommended having carrier testing for cystic fibrosis, Tay-Sachs disease, Canavan disease, and familial dysautonomia. If Steve was a carrier for any of these diseases, Vanessa would also be screened. For couples in which both partners are carriers for any of these genetic conditions, the fetus has a one in four risk of being affected, and prenatal diagnosis for the disorder is available.
Sherman told the couple what he thought the best course of action was for them, and we believe that physicians should be prepared to do this or to give a direct answer to the question “Doctor, what would you do in this situation?” The response “I’m not you, so that’s not relevant” is not useful. This is not like asking a waiter to suggest which item on the menu he or she likes best: taste in food varies widely and is not based on expertise. Patients, unlike diners dealing with waiters, seek out physicians because they are professionals and have special expertise, and it is reasonable to expect them to share their opinions about what they think is the best approach in a particular pregnancy. Of course, patients are not required to agree with their physician or to take their advice, but you will likely find your physician’s advice useful in making your own decision.
Sherman asked Vanessa and Steve to consider all the information they had discussed, but not for too long. They would have to decide what course they wanted to take within the next few days if they wanted to maximize their options, especially first-trimester screening or CVS. Sherman told them they should call him if they had any questions, and he was pretty sure they would. He also provided written information that outlined everything he had discussed with them. Perhaps most importantly, he told them that there was no right or wrong answer; their decision should be based on what is best for them.
For Vanessa and Steve, what had started out as a dream day was ending in a decision-making nightmare. When Vanessa and Steve first learned of their pregnancy, they were extremely happy and thought mostly about adding the baby to their family. They just assumed that they were going to have a healthy baby. Sherman knew that they left his office worried about whether their baby might have a serious genetic disorder or birth defect. Modern medicine had succeeded in making them much more anxious and uncertain about the health of their baby-to-be. Couples in the near future will be faced with much more genetic and genomic information about their fetus and, paradoxically, more uncertainty. George thinks Sherman did a good job of informing Vanessa and Steve of their choices, but he also thinks it is important to have another knowledgeable person, such as a genetic counselor, to consult.
Sharing genetic messages with couples is, we think, imperative. But the question of how to maximize the good that genomic medicine can do while minimizing the potential harm is one that recurs as more and more genomic information is added to the practice of medicine, and this dilemma is not unique to prenatal care. New methods will have to be developed to mitigate information overload and help physicians convey genomic messages to pregnant patients in a helpful and meaningful way. We make some suggestions later in this chapter.
Karyotypes and Chromosomal Microarrays Analysis
For the past half century, the standard cytogenetic procedure for diagnosing chromosomal abnormalities has been a microscopic analysis of a complete set of chromosomes (normally forty-six) prepared from a dividing cell (for instance, a white blood cell or amniotic fluid cell). A systematized display of pairs of chromosomes in order of decreasing size is referred to as a karyotype. If, for example, three copies of chromosome 21 are present, termed trisomy 21, it is diagnostic of Down syndrome (figure 6.1).
6.1 Karyotype for trisomy 21 or Down syndrome. Notice the three copies of chromosome 21. Wikimedia Images, June 23, 2006.
Chromosomal microarray analysis (CMA) technology detects not only entire extra or missing chromosomes but also small losses and gains of segments of DNA throughout the genome, while a karyotype analysis allows detection of extra or missing fragments of chromosomes, but only those large enough to be seen under a microscope. These losses and gains of segments of DNA are referred to as copy number variants (CNVs); CNVs can lead to genetic disorders associated with significant disabilities. CMAs are created by robots that print tiny dots arranged in a precise grid onto a glass slide. Each dot contains predetermined fragments of DNA sequences (probes) from known locations on each of the forty-six chromosomes. Microarrays typically contain millions of such dots. For prenatal diagnosis using CMAs, fetal DNA derived from amniotic fluid cells or a chorionic villi cell is “digested” (chopped up) with enzymes and then labeled with one color of fluorescent dye, while the control (“normal”) digested DNA, derived from a person or pool of people with no known genetic abnormalities, is labeled with another color of fluorescent dye. Digested and control DNA is then added to the microarray probes. If the digested DNA is complementary to the microarray probes, the fluorescent dye lights up. The microarray is then placed in a special scanner that measures the relative brightness of each fluorescent dot (that is, the relative amount of fetal DNA versus control DNA). CNVs (duplications or deletions of fetal DNA) are detected as regions with higher or lower signal strength than the control sample. CMAs can also be “targeted,” meaning that they are constructed to focus on CNVs of known pathogenicity instead of testing the entire genome.
The National Institute of Child Health and Human Development, a part of NIH, conducted a large study, called the MAStudy, to compare microarray analysis to karyotype analysis. A group of 4,406 women undergoing either CVS or amniocentesis was enrolled, and samples were split two ways: standard karyotyping was performed on one portion and microarray analysis on the other. In one in sixty cases where the karyotype was read as “normal,” microarrays revealed a CNV considered of clinical importance. When prenatal diagnosis was prompted by a structural abnormality of the fetus seen on an ultrasound and the karyotype was interpreted as “normal,” microarray analysis revealed a CNV in almost one in seventeen cases.
Clinical dilemmas are sometimes encountered when CMAs show “variants of unknown (or uncertain) significance,” or VUS. A VUS is a change in DNA that has not yet been reliably characterized as benign or pathogenic. Sometimes the answer to the question “What is the chance of this VUS leading to a significant problem in my baby?” is that we just don’t know. The vast majority of VUSes are likely to be just benign variants with no clinical consequences. On the other hand, even if a VUS is inherited from an apparently “normal” parent, the CNV may cause serious congenital and developmental abnormalities in the child. In such cases, finding a VUS could lead to questioning whether the parent who transmitted the VUS is really “normal,” or if we need to search for health problems that have not been recognized. In the MAStudy, VUSes were found in 3.8 percent of all cases where the karyotype was read as “normal.” A clinical advisory committee offered advice when VUSes were encountered. Sherman was a member of this committee. The case of Cindy and Mark (not their real names) provides an example.
Cindy, a healthy forty-year-old woman, underwent CVS during the twelfth week of her second pregnancy and participated in the MAStudy. Cindy and Mark’s only child, a daughter, was reported to be normal. The family history revealed that Cindy had a first cousin with congenital deafness. On Mark’s side, there were two first cousins, a brother and a sister, both reported to have autism and another first cousin with developmental delay and congenital heart disease. The chromosomal karyotype analysis from the chorionic villi indicated a normal female. However, CMA showed loss of a segment of DNA in one of the two chromosome 16s in the fetus. Cindy and Mark underwent CMA studies to see if one of them carried the loss of the same DNA segment, and it turned out that Cindy did.
The clinical advisory committee was consulted because similar losses of DNA segments on chromosome 16 have been found in both normal individuals and individuals with mental retardation, autism, seizures, and schizophrenia. An unaffected parent can have an affected child, even when both carry the same deletion. The committee decided that Cindy and Mark should be informed and counseled about the CMA findings. What the couple decided to do about the finding (that is, continuation or termination of the pregnancy) was not revealed to the committee.
Twenty-three other participants in the study who had received abnormal prenatal CMA results were interviewed by genetic counselor Barbara Bernhardt. Most reported being shocked, anxious, confused, and overwhelmed after receiving the news. One woman was quoted as saying, “You know, they’re telling me there’s something wrong, but they can’t tell me what. . . . We wanted to know what that would mean for our son in the future. And they really couldn’t tell us.” Some women with uncertain results said they received conflicting information from counselors and physicians, and they also found contradictory information on the Internet. As one woman described it, “I started getting really panicky that the child I was carrying was going to be severely autistic with seizures and schizophrenia. . . . I would look online, and I met a geneticist and talked to an autism specialist. And frankly nobody could really tell me. . . . I ended up going to a crisis counselor because it was very stressful.”
The limited time available for decision making makes it imperative to gather information quickly, including testing the parents to determine whether the fetal findings are inherited. If fetal abnormalities can be seen by ultrasonography, it adds suspicion that the VUS is deleterious, but lack of such findings does not guarantee the fetus is healthy. One woman explained, “I think what is so difficult about decisions—on top of the shock of it, the helplessness—is the timing. . . . We felt like we were desperately trying to build enough information to make an informed decision in a very fast amount of time, and that was very, very stressful for us.”
Even when a CNV is known to be associated with a well-described genetic syndrome, making a decision about continuing the pregnancy may still be fraught with uncertainty. One woman, for example, terminated her pregnancy after her fetus was found to have a deletion of part of chromosome 22 (termed a 22q11.2 deletion). This deletion is associated with DiGeorge syndrome, a condition that may include heart defects, poor immune system function, a cleft palate, complications related to low levels of calcium in the blood, and behavioral disorders. Neither the woman nor the father carried the deletion. The woman said, “We still grapple with this because it is very much a spectrum of severity, very, very hard to predict what the outcome would be. . . . So that was very, very difficult for us because it made assessing our choices really hard.” Half the women whose prenatal CMA results were uninterpretable or uncertain and who delivered normal-appearing infants had regrets about having the test in the first place—called “toxic knowledge.” As one woman put it, “Since I had this uncertain microarray result. . . . If anything happens to him in the future . . . that will always pop up in my mind. . . . You just have to have a ‘wait and see’ attitude. . . . I’m a lot more vigilant.”
Genomic information possesses a mythology of precision and determinativeness that it does not deserve. As the medical literature and databases associated with CNVs expand, such diagnostic uncertainties will be less frequent, but they will still occur. While studies continue, ACOG has recommended that couples choosing chromosomal microarrays should receive both pretest and posttest genetic counseling. Prenatal counseling is usually described as “nondirective” (that is, not for or against termination of the pregnancy), but this term has no meaning in the context of a genomic VUS. Physicians must share the uncertainty of diagnosis with their patients and advise them as best they can without guaranteeing a healthy baby.
Jay Katz has argued that sharing uncertainty is particularly difficult for physicians because they cling to the notions that uncertainty is grounded in ignorance and that it is their job to replace the patient’s ignorance with their knowledge. Katz has suggested that physicians could learn from John Keats and his concept of “negative capability,” the ability to live with uncertainties, mysteries, and doubts inherent in human life. Katz argues that physicians should be like poets in that the art of medicine is akin to discovering beauty in both its life-affirming and life-destroying dimensions; the science of medicine seeks to discover truth in the “beauty of discovery and in the ugliness of ignorance.” Katz believes that dealing with uncertainty is the key to doctor-patient conversations and that medical education has neglected training in “how to truly cope with uncertainties” without “becoming paralyzed by them.” This is necessary for both physicians and patients, Katz suggests, because “it is the legacy of science that scientific activity produces not only new knowledge but also new ignorance,” and physicians need to acknowledge both to their patients. These lessons may prove particularly valuable when adopting new methods of prenatal diagnosis where, quoting Katz again, “the disregard of uncertainty defeats the sharing of the burden of decision” and promotes evasions and even lies that can make “meaningful disclosure and consent a charade.”
For more than three decades, finding a noninvasive way to analyze the genetic makeup of the fetus by identifying fetal cells in the blood of the pregnant woman has been a scientific quest because it could permit prenatal diagnosis without risk to the fetus. Sherman’s laboratory, as well as others, showed that cells from chromosomally abnormal fetuses, including trisomy 21 and trisomy 18, could be detected in maternal blood during pregnancy as early as ten weeks of gestation. An evaluation by the National Institute of Child Health and Human Development (NICHD) concluded that technological advances were needed before fetal cells in maternal blood could be used for routine clinical application. The quest to dependably collect fetal DNA from maternal blood has been taken up by private biotech companies, one of which Sherman worked with as scientific director. In 1997, it was first reported that fragments of cell-free fetal DNA (cfDNA) circulate in the maternal blood during pregnancy beginning in the early first trimester. About 10 percent of the DNA in maternal plasma is now known to be of placental origin. This offers the possibility of noninvasive prenatal screening by taking a blood sample from the pregnant woman.
The technology currently used for commercial testing is massively parallel genomic sequencing. This is a highly sensitive assay that rapidly quantifies millions of DNA fragments from maternal blood. Bioinformatician Titus Brown has likened this process to shredding one thousand copies of A Tale of Two Cities in a wood chipper and then putting them back together. ACOG considers cfDNA an option for primary screening in women at increased risk for having a child with a chromosomal disorder but does not currently recommended it for general obstetric patients. A limitation of cfDNA testing is that it provides information regarding only trisomy 21, 18, and 13, as well as the number of X and Y chromosomes, and this should be explained when counseling patients. More recent data has shown that cfDNA can be accurate in all pregnancies, not just high-risk pregnancies. Performing a cfDNA test should be an informed patient’s choice after pretest counseling and can be considered mainstream obstetrical care. This technique is being so rapidly introduced into clinical practice that it has been reported that the rate of CVS and amniocentesis has dropped significantly in the past 2 years, perhaps as much as 50 percent.
On the other hand, at least some physicians and their patients do not understand how these screening tests work, and they have aborted healthy fetuses as a result of a screening test without following it up with a diagnostic test. A story published on the front page of the Boston Globe, for example, concluded that “companies selling the most popular of these screens do not make it clear enough to patients and doctors that the results of their tests are not reliable enough to make a diagnosis.” The article features Stacie Chapman, who was told by her physician, Jayme Sloan, that the test had determined that her fetus had trisomy 18, a condition incompatible with life, and that the screening test she had, Sequenom’s MaterniT21 Plus, had a 99 percent detection rate. What neither she nor her physician appreciated was that there was a good chance the test was wrong. After discussing trisomy 18 with her husband, Stacie made the decision to terminate her pregnancy. Before she did, however, her physician called her back and urged her to have a diagnostic test, which showed that her future son did not have trisomy 18 after all. In Stacie’s words, “He is so perfect. . . . I almost terminated him.”
Other women were not so lucky and did terminate healthy fetuses on the basis of a screening test alone. Michael Greene, director of obstetrics at Massachusetts General Hospital, observed that the marketing of tests has created problems: “The companies have done a very poor job of education . . . failing to make clear that it is screening testing with very good but inevitably not perfect test performance . . . and that doctors are recommending, offering, ordering a test they do not fully understand.” Not only are some women aborting healthy fetuses, but others are giving birth to affected fetuses even after having the screening test. Belinda Boydston, for example, says she gave birth to a child with trisomy 18 even after being assured that the fetus had only a 0.01 chance of having the condition. One solution to dealing with these unexpected results is not to use these screening tests without prior genetic counseling.
The Future of Prenatal Diagnosis
Within the next few years, noninvasive prenatal detection for the majority of recognized genetic disorders will likely become a reality. The fact that the entire fetal genome is represented in maternal blood opens up the possibility of obtaining fetal DNA fragments that can be assembled into a complete fetal genome readout. A number of substantial hurdles, including the lack of genetic counseling, remain before noninvasive fetal whole-genome sequencing is introduced into clinical practice. First, the cost will have to drop substantially. In this regard, targeting selected genomic regions may prove more efficient and cost effective than deriving the whole genome. On the other hand, isolating and analyzing fetal cells from maternal blood would be more straightforward and presumably less costly. It is likely that noninvasive whole-genome sequencing of the fetus will eventually be done at a cost of less than $1,000, and sequencing the entire genome may be the most efficient approach. We are already at the point where we can identify “3,600 genes for rare Mendelian disorders, 4,000 genetic loci related to common diseases, and several hundred genes that drive cancer.”
A critical consideration for prenatal genetic diagnosis, whether by invasive or noninvasive means, is deciding which conditions warrant testing. There is little doubt that noninvasive prenatal screening will be viewed by both the public and physicians as a positive advance primarily because it avoids the risk of miscarriage produced by invasive tests. However, philosopher Evelyne Shuster has suggested that whole-genome fetal testing could act as a metaphorical “toll booth” on the road to childbirth, and that as testing for hundreds or even thousands of genetic conditions becomes feasible, it will become more and more difficult to decide if any fetus should be considered “healthy” enough to warrant carrying to term. Are we entering a new phase of eugenics—private rather than government-sponsored—with the question of who deserves to be born decided in private doctors’ offices? Is private eugenics the price we pay for private decisions? Physician-philosopher Georges Canguilhem has suggested that genomics always begins with a “dream . . . to spare innocent[s] . . . the atrocious burden of producing errors of life. At the end there are the gene police, clad in the geneticists’ science.” To the extent he is correct about government’s natural tendencies, it is our task to recognize and oppose the tendency toward a new eugenics.
Will noninvasive prenatal testing become “normalized” and routine because of its ease and safety? We think it is inevitable that reasonably priced and accurate genetic and genomic tests based on fetal cells or cfDNA extracted from maternal blood will be offered to all pregnant women. The primary motivation, however, will likely not be medical practice standards but fear of medical malpractice lawsuits. Obstetricians will likely fear a malpractice lawsuit if the fetus is born with a genetic abnormality for which the couple can credibly explain to a jury that they would have had the genetic test (had they known about it) and terminated the pregnancy. We both believe that physicians’ fear of the extremely unlikely prospect of such a lawsuit is a terrible way to set medical practice standards. Rather, we support professional organizations that set screening standards for their members—and whose members follow them.
Not all screening and testing decisions, even for fetuses, will be made in the context of the doctor-patient relationship. Aggressive marketing and consumer demand will likely drive the utilization and adoption of new genomic technologies. We are already seeing cfDNA noninvasive prenatal screening companies reaching out to pregnant “consumers” on YouTube, Facebook, and Twitter. Companies that can consistently predict fetal sex early in pregnancy seem likely to find a consumer market in the United States. A market is also likely to exist for noninvasive prenatal paternity testing. Some women may decide that without paternity testing they would terminate the pregnancy because they could not be sure of the identity of the “right father” for the fetus, a story line at the center of the second season of the popular zombie series, The Walking Dead.
Translating and conveying expanding genomic information into useful knowledge is becoming increasingly challenging. Taking a family history and, when indicated, recommending tests to identify carriers of genetic diseases are now standard in obstetric care. Today’s carrier screening tests usually focus on conditions that occur in the ethnic groups of one or both prospective parents. With noninvasive prenatal testing, we foresee a time when, rather than screening parents to identify carrier status for genetic disorders, fetuses will be directly tested to determine whether they are affected. (This would allow fetal testing without determining paternity.) It will then become increasingly difficult—if not impossible—to inform those offered screening or testing for reproductive purposes about all the genetic and genomic information that can be obtained and the implications of that information.
Our current model for prenatal screening and diagnostic testing requires pretest counseling prior to obtaining informed consent, and the obligation to counsel can be seen as inherent in the fiduciary nature of the doctor-patient relationship. For ordinary medical procedures, the physical risks and treatment alternatives (those things that might lead a patient to reject therapy or choose an alternative) are the primary items of information that must be disclosed and should be discussed. Self-determination and rational decision making are the central values protected by informed consent. In the setting of reproductive genetics, what is at stake is the right to decide whether or not to have genomic testing, with emphasis on the right to refuse if the potential harm (in terms of stigma or unacceptable choices, including abortion) outweighs, for the individual or family, the potential benefit.
Hundreds of new genomic screening and diagnostic tests, including exome and whole-genome sequencing, addressed in the following chapter, will compete for introduction into routine clinical practice. As we outlined at the beginning of the chapter, critical questions include these: What information should be provided to which patients? When should it be provided? How and by whom should it be conveyed? It will soon be impossible to do meaningful counseling about all available genomic testing. Giving too much information (information overload) can amount to misinformation and make the entire counseling process misleading or meaningless.
To prevent disclosure from being pointless or counterproductive, we believe that information-sharing strategies based on general or “generic” consent should be developed for genetic and genomic screening and diagnostic testing. Their aim would be to provide sufficient information to permit patients to make informed decisions yet avoid the information overload that could lead to “misinformed consent.”
Traditionally, the goals of reproductive genetic counseling are to help the person or family to
• comprehend the medical facts, including the diagnosis, the probable course of the disorder, and the available management choices;
• appreciate the way heredity contributes to the disorder and the risk of recurrence in specified relatives;
• understand the options for dealing with the risk of recurrence;
• choose the course of action that is appropriate in view of their risk and their family goals, and act in accordance with that decision; and
• make the best possible adjustment to the disorder in an affected family member or to the risk of recurrence of the disorder.
As we saw with Vanessa and Steve, even knowledgeable couples can become confused, frustrated, and anxious if faced with multiple options for genetic screening and testing. An approach based on “generic consent,” an approach we described in the New England Journal of Medicine twenty years ago and still believe in, would not even attempt to describe each of hundreds or thousands of genetic conditions and anomalies to be screened and tested for, but would instead emphasize broader concepts and common-denominator issues in genetic and genomic screening.
We envision a doctor-patient relationship in which patients are told of the availability of a panel of genetic and genomic tests that can be performed on a single blood sample, either for carrier screening or noninvasive prenatal testing. Couples would be told that these tests focus on disorders that involve serious physical abnormalities, mental disabilities, or both. Several common examples would be given to indicate the frequency and spectrum of severity of each type or category of genetic condition for which screening or diagnostic testing was offered. Conditions such as spina bifida and cleft lip, chromosome abnormalities such as Down syndrome and trisomy 18, and single-gene disorders such as cystic fibrosis and Tay-Sachs disease, might be chosen as representative examples.
In the course of counseling, important factors common to all prenatal screening and diagnostic tests would be highlighted. Among these are their limitations, especially the fact that negative results cannot guarantee a healthy infant. For screening tests, the couple needs to know that additional, invasive tests may be needed to establish a diagnosis or clarify confusing or uncertain results. Other considerations that need to be discussed are options (such as adoption, egg or sperm donation, abortion, or acceptance of risks); the costs of testing; and issues of confidentiality, including potential disclosure of the results to other family members. If the testing is for carrier status, and a recessive gene is detected in the woman, it must be emphasized that her partner should also be screened (and this could bring up the issue of nonpaternity).
For prenatal diagnosis, couples should understand that abortion of an abnormal fetus is available but not the only option. For couples for whom abortion is unacceptable for any indication, prenatal diagnosis may still provide important information. For example, if it is known that a baby will be delivered with serious birth defects, choosing delivery at a tertiary medical center that offers specialized care may optimize the infant’s outcome. In some cases, knowing that the fetus has a very serious and incurable disorder may alter obstetrical care for the mother. For example, if the fetus has trisomy 18, there is a high likelihood that fetal heart monitoring during labor would show an abnormal tracing. Knowing that the fetus has trisomy 18 could avert performing a cesarean delivery because it would not benefit the infant.
Generic consent to genomic screening and diagnostic testing can be compared with obtaining consent to perform a routine physical examination. Patients know that the purpose of the examination is to locate potential problems (the doctor is looking for trouble, and the patient is hoping that no trouble will be found) that are likely to require additional follow-up and that could present them with choices they would rather not have to make. The patient is not generally told, however, about all the possible abnormalities that can be detected by a routine physical examination or routine blood work, only the general purpose of each. On the other hand, tests that may produce especially sensitive and stigmatizing information, such as screening blood for the human immunodeficiency virus (HIV), should not be performed without specific consent. Similarly, because of its reproductive implications, genetic testing has not traditionally been carried out without specific consent.
What is central in generic consent for genomic screening and diagnostic testing is not a waiver of the individual patient’s right to information. Rather, it would reflect the physician’s fiduciary obligations to the patient and a conclusion that the most reasonable way to conduct genomic screening and diagnostic testing for multiple diseases is to provide basic, general information to obtain consent for the screening, and much more detailed information on specific conditions if they are detected. Since, in the vast majority of cases, no such conditions will in fact be found, this method is also the most efficient and cost effective.
Some patients may want more specific and in-depth information on which to base their decision regarding testing. It is therefore essential to build into the consent process ample opportunity for patients to obtain all the additional information they want to help them make decisions. Clinicians, of course, must be open and responsive to the concerns and questions of patients. Counseling could be provided in person by a physician or other qualified health professional. Alternatively, Internet-based audiovisual aids could be used, which could help ensure consistency in the information provided, improve efficiency, and respond to the shortage of genetic counselors.
Generic consent for genetic and genomic screening and diagnostic testing should help prevent overloading patients with information and wasting time on useless information, especially for carrier screening. It would not, however, solve what is likely to be an even more central problem in prenatal genetic testing: are there genetic conditions for which testing should not be offered to prospective parents? Examples might include genes that predispose a person to a particular disease that will not appear until late in life (such as Alzheimer disease, Parkinson disease, or breast cancer). From the perspective of the fetus, life with the possibility—or even a high probability—of developing these diseases in late adulthood is much to be preferred to no life at all. Thus, in this case, unlike that of the fetus with trisomy 18, for example, no reasonable argument could be made that precluding abortion by denying this information could amount to forcing a “wrongful life” on the child.
We should, nonetheless, directly and publicly address the question of whether there are conditions for which screening prospective parents or testing fetuses should not be offered as a matter of good medical practice and public policy, regardless of the technical ability to do such testing and the wishes of the couple. Offering genetic and genomic screening and diagnostic testing to assist couples in making reproductive decisions is not a neutral activity, but rather implies that some action should be taken on the basis of the results of the test. Simply offering carrier screening for breast cancer or colon cancer genes in the context of preconception care, for example, suggests to couples that artificial insemination, adoption, and even abortion are all reasonable choices if they are found to be carriers of such genes. On the other hand, because of a personal experience with a family member who suffered from one of these adult-onset diseases, a particular couple might see abortion as a reasonable choice under such circumstances. A practice of keeping such information away from all couples would not be justified. However, in general we do not believe that pregnancies in women who want to have a child should be terminated for adult-onset diseases. We are all going to die of something, and if we live long enough, that something will have a major genetic component. There can be no perfect genome, and the search for it in a fetus will inevitably fail.
A standard of care for genomic screening and diagnostic testing, as well as a standard for informed consent in the face of hundreds or thousands of available genomic tests, will inevitably be set. We believe the medical profession should take the lead in setting such standards and that, with significant public input (not common today) and support, the model of generic consent for genomic screening and diagnostic testing will ultimately be accepted.
After you finish this chapter, it may seem that there is no such thing as a “healthy” fetus. The practice of prenatal screening and diagnosis is founded on looking for problems. However, with the rapidly increasing number of genomic and other types of testing during pregnancy, it is increasingly likely that one or more “abnormal” results will turn up, at least some of which may be “serious.” Given all this anxiety in both perspective parents and their physicians, it may seem remarkable that the vast majority of babies are just fine. Perhaps the hardest lesson for us to accept is that the future is uncertain. We all want reassurances that our future, especially the future of our loved ones, including our future children, will be all right.
The future is unknowable, and there really are no guarantees, no matter how much testing is done. This leads to stress and anxiety, and we must live with ambiguity. This is the price we pay for the information we receive from testing and the opportunities it gives to make potentially better choices for our lives. Keats can help us here, as he helped Katz with his idea of negative capability. In the most anthologized poem in the English language, To Autumn,Keats accepts mortality but finds beauty and truth in aging: “Where are the songs of Spring? Ay, where are they? Think not of them, thou hast thy music too.” We are in the early spring of prenatal genetic screening, and if done right, future generations will look back on these times in genomics as the “songs of Spring” that opened a conversation we will think of with fondness and pride.
In the next chapter, we present even more genomic screening tools that are used to look for specific problems—but this time problems that are treatable if detected in a child or newborn. In the context of newborn screening, we again address the question of whether and how expanding the number of genomic screening tests can reasonably be introduced into medical and public health practice. As you have probably already guessed, although we can divide genomic screening by populations, including adults, children, newborns, and fetuses (and even preimplantation embryos, as discussed in the previous chapter), once genomic screening is viewed as reasonable for any of these populations, it is almost inevitable that at least some physicians and patients will uncritically see it as reasonable for all of them.
WHEN THINKING ABOUT
PRENATAL GENOMIC SCREENING,
CONSIDER THESE THOUGHTS
Prenatal screening has relied primarily on
measurement of proteins in the pregnant
woman’s blood but is fast moving into cell-free
placental DNA in the pregnant woman’s blood.
We may soon be able to use cell-free DNA
from maternal blood to do noninvasive
whole-genome sequencing of the fetus.
There is no perfect genome and no
genetically perfect fetus.
Informed consent is fostered by physicians
sharing uncertainty with patients.
New consent models for prenatal screening will
have to be developed and should be judged on
whether they improve the patient’s right to decide.