Genomic Messages: How the Evolving Science of Genetics Affects Our Health, Families, and Future - George Annas, Sherman Elias (2015)
Chapter 7. Genomic Messages from Newborns and Children
There is no difference between men, in
intelligence or race, so profound as the difference
between the sick and the well.
—F. Scott Fitzgerald, The Great Gatsby (1925)
There are two primary ways to use genomics for our children. The first is to attempt to diagnose a child whose illness cannot be diagnosed by conventional medical means. There are a handful of famous cases in which children who were very sick with unidentifiable, rare conditions were successfully diagnosed and treated using whole-genome sequencing, and more are continually being added to the list. The second, is to use genomic sequencing at birth or during childhood to screen healthy children for treatable conditions. In this chapter, we examine both these uses of genomics. In the first we will examine a classic challenge in medicine: using our vastly increased genomic information to diagnose a rare childhood disorder. In the second, newborn screening, we address what to screen newborns for, including the complex role of the state in mandating screening and the role of the parents in providing informed consent to genomic screening. As we will see, because newborn genomic screening can detect genetic risks that will not be relevant until adulthood, children have a dignity interest in making their own decisions about learning this information when they become adults. Likewise, because the genetic information of the newborn also contains genetic information about the parents, the parents have a personal health interest beyond the interest they have in their child’s health.
Next-generation sequencing encompasses a variety of technologies that enable rapid sequencing of many large segments of DNA, up to and including entire genomes. A distinction must be made between whole-exome sequencing (WES) and whole-genome sequencing (WGS). WES is sequencing the “exome,” that portion of the genome that codes for proteins. The exome makes up only about 1 percent of the genome, but it is the part most likely to include mutations that increase the risk of diseases. WGS implies sequencing most or all of the DNA content (protein-coding as well as the remainder of the genome), although there may be components of the genome that are not included in a present-day “whole-genome sequence.” Costs increase as one goes from WES to WGS, though the difference will decrease over time. We use the term genome sequencing to include both WES and WGS. As the cost of sequencing goes down and the number of potentially useful discrete genetic tests increases, a tipping point will be reached at which obtaining data by genomic sequencing will be more efficient than looking for mutations in individual genes. This change will be reflected in all applications of genomic sequencing in clinical medicine, including newborn screening. Of course, costs cannot control use: if a test has no practical clinical use or causes more problems than it solves, it should not be used even if it is free.
Genome Sequencing of Ill Children
Genome sequencing is not reasonable for healthy children but may be appropriate for a particular ill infant or child. The ACMG states that WES/WGS should be considered if a patient is believed to have a likely genetic disorder, but specific genetic tests available for the clinical picture have failed to arrive at a diagnosis. The ACMG and American Academy of Pediatrics (AAP) recommend that parents be informed of potential benefits and potential harms, and their consent obtained. Of course, whether these groups recommend this or not, informed consent is a legal and ethical requirement. The benefits include determining possible preventive or therapeutic interventions, decisions about surveillance, clarification of the diagnosis and prognosis, and recurrence risks. It is also acceptable to perform pharmacogenetic testing to help with drug selection, appropriate dosing, and other safety and efficacy issues. Potential harms can occur if the results lead to pursuing unproven treatments, particularly if they are ineffective or have significant adverse effects. These recommendations properly focus on the best interests of newborns and children.
Rapid genome sequencing (in about forty-eight hours) for disease gene sleuthing has been used to provide essential information in the newborn intensive care unit (NICU). For acutely ill babies, coming up with a quick diagnosis of a genetic disorder can be vital. There are about five hundred diseases involving mutations in single genes for which there are therapies. The proper treatment can usually be determined without resorting to genomic sequencing; nonetheless, genomic sequencing can help when more conventional diagnostic tests fail to diagnose a potentially treatable condition. Unfortunately, most cases in which sequencing has been used in the NICU for rapid diagnosis have to date involved genetic disorders for which there is no effective treatment. For example, an infant girl developed seizures within an hour of being born. Despite intensive efforts to locate a possible infection and tests looking for metabolic diseases and chromosomal abnormalities, all results came up negative. A brain MRI was normal, but an electroencephalogram (a brain wave test) confirmed abnormal brain activity. Multiple antiseizure drugs were tried, to no avail. The infant had to be put on a ventilator because of low oxygen levels and a slow heart rate. At this point, whole-genome sequencing was done to attempt a diagnosis. A mutation was found in both copies of a gene on chromosome 7 designated as BRAT1. Both parents were tested and found to be carriers for the same mutation. After lengthy discussion, the parents requested the withdrawal of life support.
In cases like this one, rapid diagnosis by genomic sequencing of an untreatable and inevitably fatal disorder allows physicians and the family to stop looking for a cause. The decision can be made to move the infant to a more appropriate setting, where the parents can be given an opportunity say good-bye. Sometimes discovering that the parents are carriers for the same mutation indicates that they are at risk for having another affected child, which can be important for future reproductive decisions. Current research indicates that, using clinical genome- and exome-sequencing techniques, a causative genetic variant can be identified in approximately 25 percent of pediatric patients who have been identified as having a condition that is likely due to a single-gene genetic disorder. Because finding the genetic cause is relatively unusual and other conditions may be identified (so-called incidental or secondary findings), it is critical that realistic pretest counseling be conducted with the parents. Most importantly, even when the causative genetic variation is discovered, it is still unusual to have the diagnosis lead to a change in medical management or an improvement in prognosis. Treatment, of course, is the ultimate goal—but that goal is, for now at least, still in the distant future for most children. Two studies have been able to make a molecular diagnosis using exome sequencing in about 25 percent of sick children, and a medically actionable finding in about 5 percent of the children. Commenting on one of the studies, clinical geneticist Jonathan Berg noted that “the difficulty of interpreting the clinical significance of genomic variants” remains a challenge, and “there is much to learn before [genomic sequencing] can be applied more universally.”
Nonetheless, there have been spectacular treatment successes. Retta Beery, the mother of twins, Alexis and Noah, described her diagnostic odyssey to the President’s Bioethics Commission. Her twins were born in 1996, and it soon became clear from their jerky movements that something was seriously wrong. Over the following two years, she and her husband, Joe, took the twins to a variety of specialists, and they endured countless tests and surgeries with little result. In 2002, with a diagnosis of Segawa’s dystonia (DOPA-responsive dystonia) the doctors began a course of treatment to increase the childrens’ brain dopamine, which dramatically improved their health. In 2009, Alexis developed severe breathing problems, and in 2010 the Beerys went to Baylor University Medical Center for whole-genome sequencing. The sequencing revealed a rare and only recently recognized genetic cause of dystonia. This information enabled their neurologist to successfully treat the twins with an over-the-counter supplement.
Another highly publicized success story involves Nicholas Volker. When Nicholas was fifteen months old, his parents sought care for him because of poor weight gain, inflammation, and a large draining abscess around his anus. Antibiotics did not help, and his condition deteriorated. At two and a half years, Nicholas was sent to a major medical center, where he was found to have severe growth stunting and malnutrition. The presumed diagnosis was a severe form of Crohn’s disease, an inflammatory disease that usually affects the intestines. The cause of Crohn’s disease is not well understood, and there is no definitive diagnostic test; it is thought to result from interactions among genetic, environmental, immunological, and bacterial factors. Despite treatment, his condition continued to deteriorate. A colostomy, a surgical procedure that brings one end of the large intestine out through the abdominal wall to divert fecal material, was performed. He developed sepsis (a potentially fatal response to bacteria or other germs) and required four weeks in the intensive care unit. At age four, he was again admitted to the hospital, and his entire colon had to be removed.
Over the next several years, Nicholas continued to require hospitalizations and experienced severe, potentially lethal complications. To keep him alive, a highly aggressive and risky approach would need to be tried: a bone marrow transplant using donor cells. The success of this approach would depend on knowing the exact underlying cause of the child’s disease.
Whole-exome sequencing was performed on his DNA, identifying over 16,000 variants compared to the human genome reference. Using highly sophisticated analysis, a single substitution of a G (guanine) for an A (adenine) was found in the XIAP (X-linked inhibitor of apoptosis) gene, found on the X chromosome. The XIAP protein encoded by the XIAP gene plays a central role in the inflammation process. Further testing showed that Nicholas had a deficiency of XIAP protein that was causing his immunodeficiency disease. Based on these findings, at five years eight months, Nicholas underwent a bone marrow transplant using donor cells. Within a few months he was on a normal diet and had no recurrence of his bowel disease.
Together with the Beery twins, this case shows how powerful genomic sequencing information can be in diagnosing and successfully treating a rare disease—an example of personalized genomic medicine at its best. In the case of Nicholas, it was life saving. However, in the 2014 review article, these two cases were the only ones the authors could identify in which whole-genome screening of children led to a new treatment regime that dramatically changed the child’s clinical outcome. Of course there will likely be many more; our only point is that we are in very early days of using genomics to diagnose and treat children with rare genetically-determined diseases.
What if your child is healthy? Does it make any sense to screen him or her for “good genes” that might predict special talents? The Atlas First SportGene® Test presents this question. We discuss it in some detail because it can stand in for almost any special talent, including music, mathematics, or ballet. The company claims that the test “is geared specifically to show athletes, trainers and interested individuals where their genetic advantage lies.” The SportGene® Test focuses on a gene called ACTN3, which is involved in forceful contraction of fast-twitch muscle fibers used during sprint-type activity. In a recent meta-analysis, which combines the results of a number of studies that address a set of related research hypotheses, a variant of the ACTN3 gene called the RR genotype was found to be more common among sprint and power athletes compared to controls. Can we therefore conclude that a child found to have the RR genotype of the ACTN3 gene will be a more gifted soccer player? The answer is no. This is because an “athletic champion” is the result of multiple factors, including the combined influence of hundreds of genes that are expressed in many organs, such as skeletal muscles (fiber characteristics), lungs and blood (oxygen diffusion capacity), connective tissue (tendon stiffness), heart (maximal pump capacity), thyroid (metabolic rate), pancreas (insulin secretion and glucose control), adrenal glands (adrenalin surges), and brain and nerves (balance, reflexes, and pain tolerance), among others. Each of these individual genes may have multiple DNA variants that can result in positive or negative effects on the overall level of athletic performance. Any given gene variant may interact differently with other gene variants, so-called gene-gene interactions. In addition, environmental influences (for example, sports culture, nutrition, training intensity, coaching and instruction, temperature, and exercise equipment, to name just a few) are critical, as is just plain luck.
Variation in any single gene would likely have negligible effect on sports performance. For example, it is estimated that the ACTN3 gene accounts for only about 2 percent of variance in muscle performance. Even complete deficiency of the ACTN3 gene product, α-actin-3 protein, does not appear to preclude elite performance, as demonstrated by a Spanish Olympic long jumper with both copies of his ACTN3 gene having the X (“null”) variant. As Carl Foster, director of the Human Performance Laboratory at the University of Wisconsin – LaCrosse and coauthor of several ACTN3 studies, has said, “If you want to know if your kid is going to be fast, the best genetic test right now is a stop watch. Take him to the playground and have him race the other kids.” There will undoubtedly be companies trying to sell you tests to determine your child’s intellectual ability or musical talent. Like sports ability, there are much better tests for intelligence and musical ability: each involves directly testing the talent by actual performance.
Before moving from child to newborn screening, we should say a few words about incidental or secondary findings, genetic findings that are not what you are looking for but are discovered when doing exome or genetic sequencing (like that done for Nicholas Volker). Because screening children so rarely yields results that directly lead to more effective treatment, it has been suggested that whenever this type of screening is done, the testing lab should also look for other genes that might cause trouble to the child in the future, or for which the child’s parents might be at risk. We think this strategy is legitimate provided the parents get genetic counseling describing the nature of the added search and have the basic right of informed consent, including the right to refuse all or parts of the secondary testing. The original 2013 recommendations of the ACMG disregarded basic legal and ethical rights of informed consent, asserting that fifty-six specific genes should routinely be analyzed for mutations whenever genomic sequencing was performed on a patient. These included thirty-one genes associated with cardiovascular risks, twenty-three mutations associated with cancers and noncancerous tumors, and two genes associated with malignant hyperthermia (a muscle disorder that results in high fever and severe muscle contractions when general anesthesia is given). We strongly disagreed that a medical group could or should attempt to compromise the legal rights of their patients. We were pleased when, a year later, the college amended its recommendations to require informed consent complete with an opt-out provision for the supplementary hunt for the fifty-six genes. The new genetics, as powerful as it is likely to be in the future, is not so powerful or exceptional that it can or should change basic patient rights, including the right not to have medical tests performed without informed consent. As we put it in Science with our colleague Susan Wolf, a law professor: “Starting down the path of unconsented testing and reporting in clinical genomics leads to grave difficulties and should not be done without more careful analysis. . . . The era of medical genomics requires a trusting partnership with patients, based on respect for their rights.” We could have added that informed consent is a matter of law, not a matter to be determined by physicians, no matter how distinguished or knowledgeable they might be about genomics.
Genome sequencing of children will nonetheless likely be an easy sell to many parents, especially if it is marketed as the “Rosetta Stone” that can help ensure the best health for their children. It seems that whenever new technologies become available, we are anxious to adopt them. Recall that two fundamental characteristics of American medicine are that it is technologically-driven and individualistic. Even if we don’t know how we will use it, having access to more technology (and to more information) almost always seems to be a good thing, especially when it relates to our health. Nonetheless, unless your child has a serious condition that cannot be diagnosed without genomic screening, the right not to know is today much more important than the right to know. Almost all of the information genomics can now provide to parents of healthy children is either useless or potentially harmful. Using sequencing technologies, it is now known that every individual has 3–4 million variants. Virtually all of these variations are meaningless today. Diverting the attention of health care providers and parents to do exhaustive analysis and reporting of data for which we have no understanding, and which contains no known medical relevance, is unwarranted. There are also potential harms.
Some parents may view their child as “abnormal,” even though there is no evidence of a problem. In adults we call this the “worried well”; in the case of newborn screening, they could be called “worried parents of the well.” In this regard we agree with the conclusions of H. Gilbert Welch in his profound book Overdiagnosed: Making People Sick in the Pursuit of Health. Welch noted that although most of us have normal phenotypes, “each of us can be shown to be at high risk for some disease. So the new world of personal genetic testing has the potential to make all of us sick.” This is a significant problem, especially when it involves labeling children as sick. That is because Fitzgerald’s observation at the beginning of the chapter is correct: there really is no more profound difference in people than the difference between the sick and the well—and the last thing we should want to do is make ourselves, and our children, think we are sick when we’re not. As Welch provocatively concludes, “Ironically, the healthiest populations may be those that know nothing about their DNA.” The problem of variations of uncertain significance (VUS) is not unique to genetic testing, and there is inevitably a learning curve whenever we introduce new diagnostic tests into medical practice. For example, in magnetic resonance imaging, variants in healthy body structure are frequently encountered. Usually we do not waste time documenting such variants pixel by pixel but instead focus on those we currently understand to have clinical significance. The same should go for a VUS in the genome. Geneticist William Gregory Feero puts it another way, arguing (like Donald Rumsfeld on the weapons of mass destruction in Iraq) that there are many “known unknowns” and “unknown unknowns” that we will have to learn about before whole-genome screening can be integrated into clinical practice. The question for now is whether WGS data “will decrease uncertainty and improve outcomes or merely exponentially increase the complexity of clinical care.”
Produced almost two decades ago, the movie Gattaca still captures the uneasiness we all feel about genes in terms of destiny, discrimination, and eugenics. In Gattaca, parents using IVF and preimplantation manipulation can choose at least some of the genetic makeup of their children. To ensure that a “valid” offspring is produced, embryos are weeded out in the laboratory if they are found to have a “critical predisposition to any major inheritable disease.” There are “potentially prejudicial conditions,” such as premature baldness, myopia, alcoholism, addictive susceptibility, and propensity for violence or obesity, that can be modified at the embryo level. Those who are conceived naturally are considered “invalid.” In Gattaca, they are described as having “discrimination down to a science.”
Vincent Freeman is “invalid,” his parents having “put their faith in God’s hands rather than those of the local geneticist.” Immediately after his birth, a nurse pricks his heel. A drop of blood is inserted into an analyzing machine that prints out the baby’s health future. Antonio, his father, asks, “What’s wrong?” Vincent’s voice-over continues: “Of course, there was nothing wrong with me. Not so long ago I would have been considered a perfectly healthy, normal baby. Ten fingers, ten toes. That was all that used to matter. . . . [Now] only minutes old, the date and cause of my death was already known.” Vincent’s newborn test results are read out loud by the nurse in the delivery room: “Manic depression, 42% probability; attention deficit disorder, 89% probability; heart disorder, 99% probability; life expectancy, 33 years.”
Is Gattaca-like newborn screening in our future? Should it be? We begin with a true story about a newborn named Kimberly. Christina and Robert couldn’t have been happier. The nurse had just come into Christina’s hospital room with their new daughter, Kimberly, who had been taken to the nursery for her discharge examination. Christina noticed that Kimberly had a small adhesive bandage on her right heel and asked the nurse about it. The nurse explained that a few drops of blood had been taken from Kimberly for newborn screening for a number of rare diseases, and that it was done because “it’s the law.”
We have now left the realm of medical practice and entered into the realm of public health—where decisions are made by state governments to protect populations rather than individuals. In the context of genetics, we seem to be in the process of abandoning personalized medicine before it has even been introduced into the clinic, skipping directly to population health. This is certainly true of newborn screening, even though routine screening of newborns has had little impact on health at the population level. There are three things Christina and Robert (and all new parents) should know for sure. First, thousands of babies in the United States (out of millions) will have a very rare condition that could be detected (and hopefully treated) by newborn screening. Second, the chance of any individual baby having one of them is vanishingly small. Third, new screening tests always bring with them new worries, including those spawned by false-positive results (detecting a disease where none actually exists). See Appendix B for more on screening tests.
Mandatory public health screening laws are state laws because the states retained this arena of authority (sometimes called the “police powers”) when they delegated government power to the federal government through the Constitution. That means there can be, and is, variability among the states on both the content of newborn screening and parental consent to it. State-mandated newborn screening is performed on more than 4 million infants annually for certain genetic, endocrine, and metabolic disorders, as well as congenital hearing loss and critical congenital heart disease. The public health goal is to identify and treat conditions that could severely affect the child’s future health and even survival. In its announcement of the “Ten Great Public Health Achievements—United States, 2001–2010,” the Centers for Disease Control and Prevention (CDC) included newborn screening, pointing out that “improvements in technology and endorsement of a uniform screening panel of diseases have led to earlier lifesaving treatment and intervention for at least 3,400 additional newborns each year with selected genetic and endocrine disorders.”
Obviously newborn screening is important for those affected infants, but the odds of any particular infant being benefited is less than one in a thousand—meaning that Christina and Robert should not have to worry much about the results of the screening. As Stefan Timmermans and Mara Buchbinder conclude in their 2013 study of newborn screening in California, “Newborn screening may make a world of difference to individual families, but no available data has shown that newborn screening is associated with a reduction of infant mortality at the population level.” Newborn screening, the authors note, is a public health anomaly, since it is not aimed at common diseases: “The irony of expanded newborn screening is that the United States instituted a public health program aimed at prevention for very rare conditions.” How did this happen?
The model disease on which newborn screening was founded is phenylketonuria (PKU). It is a condition that is devastating if not diagnosed early, and the symptoms can be controlled by a restricted diet. In 1934, a Norwegian physician-chemist named Asbjørn Følling was consulted by the mother of two children, a brother and sister, with severe “feeblemindedness.” He determined that the urine of these children had very high levels of a chemical called “phenylpyruvic acid.” The condition eventually became known as phenylketonuria. In 1960, the microbiologist Robert Guthrie (who had a son and niece with developmental disabilities, the latter diagnosed with PKU) developed an inexpensive, sensitive, and simple bacterial inhibition assay, using a blood spot from the infant on special filter paper, that could be administered a few days after birth on a large-scale population basis. This test became known as “the Guthrie test.”
In 1961, President John F. Kennedy (whose sister Rosemary was mentally disabled) promised to double the money spent by the National Institutes of Health on “retardation” research, and appointed a Presidential Advisory Commission on Mental Retardation, charging it with appraising the adequacy of existing programs. The commission hired the Advertising Council, which, among other things, mounted a dramatic campaign advocating that the new PKU test “should be a must for all babies everywhere.” The National Association for Retarded Children was particularly influential in garnering political support and proposed model state legislation (because newborn screening is governed by state law) for mandatory PKU newborn screening. By 1975, forty-three states had enacted such laws and 90 percent of all newborns were being tested. Today, every newborn screening program in the United States includes PKU.
The mainstay of treatment for PKU is a strict diet with very limited intake of phenylalanine, which is mostly found in protein-rich foods. The goal is to consume only the amount of phenylalanine required for normal growth and body processes but no more. Because breast milk and regular infant formula contain phenylalanine, babies with PKU must have their diets substituted with a phenylalanine-free infant formula. It was once believed that it was safe for a person with PKU to stop the diet in adolescence, but it is now recommended that patients remain on a phenylalanine-restricted diet for life. On the other hand, long-term data is very spotty, with almost 70 percent of PKU patients lost to follow-up.
In addition to deciding which specific diseases warrant inclusion in screening panels, the major ethical issue with newborn screening is the informed consent of the parents. The vast majority of states make newborn screening mandatory, although (like childhood vaccinations) most have ways to opt out for religious or other reasons. Of course, you cannot opt out if you don’t know about it in the first place. In 1982 Ruth Faden and her colleagues published a study of informed consent from Maryland, a state that has required informed consent for PKU screening since 1976. Their primary question was whether the parental consent requirement interfered with the public health mission of the screening. The conclusion was that it did not. In fact, of the 50,000 women studied, only 27 (or about 1 in 2,000) refused to consent, a number, the authors note, that is 100 times less than the chance of missing a PKU infant because of a false-negative result. In a companion article, the authors argue that consent should not be required because the central issue is child welfare, not parental autonomy. In the context of PKU screening, they ask, “Is a public policy that grants parents the right to consign their children to a state of irreversible mental retardation morally acceptable? We think not.”
George took the other side, arguing that Faden’s own data undermined her ethical argument. First, there really is no conflict here between the autonomy of parents making their own decisions (liberty) and the requirement that parents act in the best interests of their children (beneficence). With such a low rate of refusal, we can accommodate both values simultaneously. As George noted, at the observed rate of refusals (1 in 2,000), “it would take 500 years before one case of PKU is missed” because of parental refusals.
Second, parental refusal cannot in any reasonable way be construed as child neglect or denying your child an important medical benefit. This is because we know to a higher degree of certainty than we know about most other things in medicine that your child does not have PKU. This is because only one in 15,000 infants has this condition, making the odds “overwhelming that if any particular set of parents refuse screening, no detriment at all will befall their child.” The real question is not about the parents but about the state: “Is mandatory screening for PKU a legitimate exercise of the state’s public health powers?” The answer to that question “requires an analysis of testing and treatment technologies, the incidence of the disease, resource allocation, and the role of law in promoting the nation’s genetic well-being.”
George agreed that newborns should be tested for PKU based on these criteria (and so does Sherman) but argued that the Faden study demonstrated mandatory screening laws to be unnecessary to protect children. George also worried about the future of screening for hundreds of conditions, which would inevitably yield high numbers of false positives. George noted specifically that if we screen for one thousand diseases and each test is so good it has only a 1 percent rate of false positives, each infant will initially be diagnosed with ten diseases—even though the infant in fact has none of these—likely generating pathology from the retesting procedures and making refusal of testing in the first place a rational decision.
We have not yet gotten to that thousand-disease test threshold, but we are moving in that direction. Driven primarily by new technology that permits rapid assessment of many conditions via tandem mass spectrometry, in 2006 an ACMG task force specifically recommended that all state-based newborn screening programs adopt an expanded uniform panel of twenty-nine core (primary) conditions and twenty-five secondary conditions. Subsequently, two additional conditions were added to this list, severe combined immunodeficiency (SCID) and critical congenital heart disease (CCHD), for a total of thirty-one primary conditions.
All states now screen newborns for the recommended twenty-nine core conditions, and some now include one or both of the newest core conditions, SCID and CCHD. The number of secondary conditions included in newborn screening panels varies among states, as do requirements for parental education about newborn screening, consent and notification processes, payment for newborn screening, provision of care for affected children, laboratory standards, and storage, use, and disposal of blood specimens. For example, coverage for medical formulas needed to meet a child’s metabolic requirements and foods needed for dietary treatment is a patchwork across states. States also finance their newborn screening programs in different ways. Most states collect a fee for screening, which differs depending on the disorders in the testing panel. For state-required conditions, private insurance or Medicaid usually covers the costs. Parents can also pay private laboratories for additional conditions, although there is almost never any reason to do this.
A number of professional organizations and government agencies have published guidelines for newborn screening based on three main considerations: (1) there should be a health benefit to the child from detecting the disorder in the newborn period, (2) the overall benefits and risks should be balanced by the financial costs, and (3) the harms should be minimal as measured by false assignment of positive or negative results. All of the guidelines emphasize that newborn screening is not just a blood test. It is a process that involves communicating timely information to parents, sampling, quality laboratory analysis, counseling when necessary, appropriate referral of an affected baby for the start of treatment, access to treatment, and follow-up evaluation of outcomes. The primary rationale for including a condition in a uniform screening panel is that it has a direct benefit to the newborn child.
Newborn screening has undeniable benefits, but only for a very small number of people. For example, early diagnosis of PKU and initiation and maintenance of dietary treatment can help prevent severe, irreversible brain damage and intellectual disability. In the United States, about two hundred babies born with PKU are detected annually. But, as already suggested, there are real downsides associated with newborn screening. Perhaps the most vexing problem is a false positive, when an infant is identified as needing additional testing because the screening results are outside the normal range, but follow-up testing shows the infant is actually unaffected. A minimumestimate is that the false-positive rate for newborn screening is 1 in 300 infants. Assuming a U.S. birth rate of 4 million per year, this would translate into more than 13,000 infants having false-positive results. One study found that, on average, there are more than 50 false-positive results for every true-positive result identified through newborn screening in the United States. For any particular test, the rate may be even higher. For example, a study of maple syrup urine disease found that there were sixty-eight false positives for every true positive.
Johanna L. Schmidt and her colleagues studied the impact of false-positive newborn screening results on families. They interviewed parents with children whose follow-up test results indicated that the initial newborn screening result was a false positive. The stress, adverse parent-child relationships, and increased perception of child vulnerability these parents experienced are evident from three representative mothers:
Upon receiving the initial positive newborn screening result, a mother (Helen) said, “I’m starting to get emotional here, but it was very stressful for me. . . . I was thinking, ‘Oh, my gosh, is he going to die?’”
During five follow-up retesting procedures of her child, a mother (Yolanda) described her experience: “The state kept on saying, you know, ‘Okay, his levels were low, but they weren’t low enough, come back.’ So much that when [my baby] saw the nurses, he would just cry.”
Another mother (Katelyn) described her feelings during the follow-up testing: “So, when she pricked my baby’s feet . . . it looked so painful. . . . He started to cry, and it’s like, ‘Oh my goodness, I’m just torturing my child, just for this.’”
For most parents, the anxiety they experienced from the time they learned the initial screening results continued until they learned that it was a false positive. One mother (Margaret) described the period as “a nightmarish two and a half weeks.” A second mother (Elizabeth) described how nothing could calm her worry: “Nothing made me feel better. . . . If God had come down, if a prophet had come down from heaven and told me that he did not have cystic fibrosis, I would not have believed it until I got the sweat test.”
Related to the problem of false-positive results is overdiagnosis, which occurs when confirmatory testing of a child with a positive screening result shows a biochemical abnormality or a gene mutation identical to that of a child with the disease, but the child never develops symptoms. When a physician encounters an asymptomatic child with abnormal laboratory results consistent with a disease, there is the dilemma of whether or not to proceed with treatment, especially if the disease is one that is known to sometimes, but not always, be life threatening or cause permanent harm. An example is 3-methylcrotonyl-coenzyme A carboxylase deficiency, which has outcomes ranging from asymptomatic (that is, a nondisease) to death.
An in-between situation occurs when an infant undergoes confirmatory testing (due to positive newborn screening results), and the follow-up tests are inconclusive. For example, the initial newborn screening for cystic fibrosis (CF) involves examining the dried blood filter paper for increased levels of immunoreactive trypsinogen (IRT), an enzyme produced by the pancreas. If positive, the next step is to look for mutations in the CFTR gene that result in an abnormal protein, which causes the clinical symptoms of cystic fibrosis. However, not all CFTR mutations cause disease, and clinical testing that entails measurement of sweat chloride values is not always clearly in the diagnostic range. Older individuals in these categories have been characterized as having “atypical cystic fibrosis” or “mild variant cystic fibrosis,” but they present for diagnostic evaluation because of signs or symptoms, whereas infants identified by CF newborn screening are symptom free. For infants identified through newborn screening programs in which cystic fibrosis cannot be diagnosed or clearly ruled out, the term CFTR-related metabolic syndromehas been proposed. Whether these children will ultimately develop cystic fibrosis–like symptoms is uncertain. This inevitably results in numerous follow-up tests and “watchful waiting,” where treatment is not given unless symptoms appear or change. This is stressful for the child and parents, as they wait for the possibility that the disease may appear.
Parental education about newborn screening needs significant improvement. Only about half of U.S. newborn screening programs provide parents with educational materials such as pamphlets or brochures, and content is inconsistent. If information about newborn screening is provided, it is usually given in the hospital shortly after delivery, when the parents are often emotionally and physically exhausted and focused on many other things. The American College of Obstetricians and Gynecologists sensibly recommends that education about newborn screening be provided during prenatal visits.
Genomic Sequencing of Newborns
Newborn screening controversies were highlighted in a report by the President’s Bioethics Council in 2008, which concluded that the concept of benefit was being greatly and uncritically widened to include not just the child but the family and society itself. Reflecting on a future when genetic screening might be available for newborns, the council argued that harms involved in detecting disease susceptibility and diseases for which there is no treatment will be accentuated and not balanced by benefits: “Today’s debates over whether to add this or that rare disorder to a uniform screening panel will be swamped, in the context of genomic medicine, by a radically more expansive approach to genetic screening.”
Should we begin to think about replacing current newborn screening (largely based on the use of tandem mass spectrometry to measure various individual chemical compounds in the blood) with whole-genome sequencing? Since the underlying basis for these diseases almost always involves gene mutations, the question must be faced: would it not be better to have the entire genome sequence and thereby detect mutations in all genes at birth—or, as envisioned in Gattaca, even before pregnancy by preimplantation genetic diagnosis—rather than performing DNA analysis on a single gene or a series of genes each time a diagnosis is sought? Whole-genome screening would permit detection of diseases even rarer than those now on current newborn screening panels. The individual’s genome sequence readout could then be entered once into his or her electronic health record and queried as needed for diagnoses, treatments, and prevention measures for the rest of the child’s life. We must debate whether this would be a good thing, but like tandem mass spectrometry in the newborn screening arena, genomic sequencing technology and its cost will likely be the primary drivers in our technology-driven health care system.
The major reason for our arguably gloomy view is economics. Routine genome sequencing of potentially every baby born in developed nations provides a massive market for companies competing in this arena, and many companies will soon begin to market their WGS tests to countries such as China and individual states in the United States. Nonetheless, states may be able to resist the sales pitches, since the cost of follow-up will be many times the cost of the screening itself, and controversy already exists over what samples and what data of the newborn the state should be able to retain and use. Moreover, having the entire genome sequence of a newborn (or adult, for that matter) will likely not be all that valuable in terms of disease prediction, treatment, or prevention for individuals, at least not for the foreseeable future. There will be exceptions, such as finding mutations for serious conditions, including inherited cancer genes (a topic we cover in the next chapter). The critical issue is whether finding a variation (mutation) in the genetic code is medically actionable; in other words, does it translate into a clinically relevant action that can cure, ameliorate, or prevent a disease or other adverse impacts on health?
The data that will become available from genome sequencing is essentially endless, akin to the amount of information one can access on the Internet, and like the Internet, only a small portion of the information will be of value to any individual. Who should decide what we look for and what we disclose, and how, especially from newborn screening? The easy answers—search for everything but tell the parents nothing, or tell the parents everything—are both equally wrong. The challenge will be to find a reasonable middle ground, and this is unlikely unless we carefully consider the ethical, legal, and practical issues now.
Bioethicists Karen Rothenberg and Lynn Bush write plays as a pedagogical approach to genetics. The fictional characters in one suggest a possible future conversation on newborn screening: “Dr. Labgen offers with enthusiasm, ‘Newborn whole-genome sequencing will help save more kids.’ Having a different opinion, Dr. Pedethic counters, ‘You’ve got to be kidding me. From what? There is no way whole-genome sequencing of newborns won’t create a massive degree of patients in waiting.’” The committee chair concludes, “What we really need is good empirical data . . . where, when, and for whom it will help, and under what conditions it would cause more harm than good.”
Anticipating this debate, and perhaps seeing routine newborn genomic screening as an potential grand entry way to introduce genomics as a central element of clinical use, NIH announced in late 2013 that it had funded four research projects on using WGS on newborns. Robert C. Green and Alan Beggs of Boston Children’s Hospital lead one of them. Green’s part involves recruiting the parents of healthy newborns who agree to have their child’s genome sequenced, then providing information to the parents to see how they use it. It will not be possible to assess the outcomes for at least eighteen to twenty years, arguably longer. So we think the newborn researchers are very premature in concluding that “we are entering an era where all of medicine is genomic medicine.” We think it is much more accurate and evidence based to say that “some” medicine will be genomic medicine, and it will not necessarily include newborn screening. We support research, but research on children who cannot give their consent can be done only with the meaningful consent of their parents, and parents can only legitimately provide consent if the risk to their children is minimal and the children can be protected from predictable harm, including stigma. The challenges of doing genomic research on healthy newborns will be to protect the child’s right to an open future, and avoiding an early genomic designation that puts the child in the “sick” category for life.
We also agree with the assessment of the editors of Nature that newborn WGS research highlights five questions:
1. Should it be done when we don’t know “with any certainty what a given genetic variant will mean for a given individual”?
2. Is it likely that providing parents with uncertain information, such as cancer risks, will lead to “years worrying about that cancer risk in their perfectly healthy child”?
3. Can clinical sequencing provide useful and timely information to parents?
4. “Who owns the genetic data?” (a topic we take up in chapter 9).
5. “Should the data be shared with other researchers?”
We also think the Nature editors were right to place ethics before science in their conclusion that when dealing with newborns, it is imperative that scientists get the ethics and science right” (emphasis added). We do not believe that state-mandated genomic sequencing of newborns is either desirable or inevitable. We should not seriously consider adopting even private genomic sequencing of newborns before it can be demonstrated, based on valid research, that it is likely to do much more good than harm, and that parents are capable of providing informed consent for themselves and their children.
Newborn screening has been sold based on the model of PKU disease. Public health in the United States strongly supports medical screening that might help even a few people. At the same time, American medicine tends to marginalize methods that could actually greatly improve the overall health of the large group of extremely premature infants in the United States. We rank very poorly in the world in terms of infant mortality, and the concentration on genetic screening of newborns will have no impact on this sad state of affairs. Nonetheless, we love our technology, and genetic screening, whether based on evidence or not, is quickly moving from looking for biochemical markers for newborn diseases to broadening out to the entire population to search for genetic markers for cancer. Cancer sells, and it is to cancer that we turn in the next chapter.
WHEN THINKING ABOUT GENOMIC
SCREENING OF CHILDREN AND
NEWBORNS, CONSIDER THESE THOUGHTS
WGS of children is in its infancy and is
only medically indicated for children with
serious, undiagnosed conditions.
WGS of sick children and newborns should
never be done without parental consent.
Newborn screening is a public health
tool for population screening.
WGS of healthy children and newborns should
not even be seriously considered until research
demonstrates that it is likely to produce
much more benefit than harm to children.
Labeling a child genetically abnormal can
transform a healthy child into a sick child for life.