From the late 1960s until recently, there was a durable consensus on the ethical principles that ought to guide the practice of newborn screening. The most in-depth and influential presentation of those principles was the 1968 World Health Organization monograph by James Wilson and Gunnar Jungner, Principles and Practice of Screening for Disease.1 The ten Wilson-Jungner criteria for including a condition in a screening program are as follows:

1. The condition sought should be an important health problem.

2. There should be an accepted treatment for patients with recognized disease.

3. Facilities for diagnosis and treatment should be available.

4. There should be a recognizable latent or early symptomatic stage.

5. There should be a suitable test or examination.

6. The test should be acceptable to the population.

7. The natural history of the condition, including development from latent to declared disease, should be adequately understood.

8. There should be an agreed policy on whom to treat as patients.

9. The cost of case-finding2 (including diagnosis and treatment of patients diagnosed) should be economically balanced in relation to possible expenditure on medical care as a whole.

10. Case-finding should be a continuing process and not a “once and for all” project.3

Wilson and Jungner emphasize the crucial importance of their second criterion: “of all the criteria that a screening test should fulfill, the ability to treat the condition adequately, when discovered, is perhaps the most important.”4 They offer an exceptionally cogent explanation of this principle:

In adhering to the principle of avoiding harm to the patient at all costs (the primum non nocere of Hippocrates), treatment must be the first aim. For declared disease there is, of course, the ethical obligation to provide an accepted treatment whether or not this is of scientifically proved value; but, when new territory is being explored by the early detection of disease, it is clearly vital to determine by experimental surveys whether a better prognosis is given by treating the conditions found at an earlier stage than was previously the practice. Unless this is so, there can be no advantage to the patient and, in fact, in alerting him or her to a condition that has not been shown to benefit by treatment at an earlier stage actual harm may be done.5

The third Wilson-Jungner criterion—the availability of facilities for diagnosis and treatment—means that a large-scale screening program cannot be justified unless resources will be available both to confirm the diagnosis and to treat effectively those who are identified as having the disease. As we shall see, this can be an issue in state-mandated newborn screening if states have the resources to screen newborns for a condition but not necessarily to follow-up and manage the care of all those who test positive.

The seventh Wilson-Jungner criterion—that there must be an adequate understanding of the natural history of the disease —is also highly pertinent to the subject of this white paper. Wilson and Jungner suggest certain questions that need to be answered about a medical condition before screening can be justified:

• What changes should be regarded as pathological and what should be considered physiological variations?
• Are early pathological changes progressive?
• Is there an effective treatment that can be shown either to halt or to reverse the early pathological condition?6

Wilson and Jungner point out the risks of embracing population-wide screening and treatment before a disease is well-understood and before controlled clinical trials have established the effectiveness and benefits of intervention:

Without well-planned surveys, carried out in advance of the main body of medical opinion, the view that early diagnosis and treatment successfully improves the outlook for the condition in question is likely to become generally accepted. This in turn automatically renders unethical planned randomized trials of intervention by treatment, following early diagnosis; with the result that ideas about the effect of treatment pass into the realm of folklore rather than of scientific knowledge.7

Although not specifically formulated for pediatric screening, the Wilson-Jungner criteria have largely guided the practice of newborn screening over the past four decades. Thus, in an important 1974 paper on the principles that should govern pediatric screening,8 pediatrician William K. Frankenburg presented those principles in a way that clearly echoed the work of Wilson and Jungner:

The availability of a suitable screening test does not justify screening for a disease unless the disease is important, relatively prevalent, and amenable to early treatment. Screening for a disease which has the necessary characteristics cannot be justified unless there is an acceptable, reliable and valid test which can be carried out at reasonable cost.

Screening which is carried out without knowledge and consideration of these criteria is likely to be wasteful of scarce medical resources and may actually do more harm than good.9

Similarly, a 1994 report of the Institute of Medicine (IOM, one of the four U.S. National Academies) on assessing genetic risks recommended that “newborn screening only take place 1) for conditions for which there are indications of clear benefit to the newborn, 2) when a system is in place for confirmatory diagnosis, and 3) when treatment and follow-up are available for affected newborns.”10 In 1995, the American Society of Human Genetics (ASHG) and the ACMG issued a joint report affirming that “timely medical benefit to the child should be the primary justification for genetic testing in children and adolescents,”11 and this judgment was reaffirmed by a 1997 report by the NIH Task Force on Genetic Testing.12 In 2000, a report by the American Academy of Pediatrics stated that a condition is a good candidate for newborn screening only if “the treatment for the condition is effective when initiated early, accepted among health care professionals, and available to all screened newborns.”13 In all these statements of principle, direct benefit to the newborn child was identified as the paramount and indispensable criterion for inclusion of a disease in a uniform screening panel. To justify such inclusion, the natural history of the disease must be well understood, the diagnostic test for its presence must be clear and precise, and an effective treatment must be available.

Despite this enduring consensus, the principle “screen only if you can effectively treat” has not gone unchallenged. A 1975 report by a committee of the National Research Council (NRC, another of the four National Academies) began by stating that newborn screening is appropriate when there is evidence that it provides “substantial public benefit,” i.e., benefit not limited to the timely and effective treatment of the infant's condition. The report went on to describe three forms of such benefit other than direct treatment: 1) to the infant (to provide management and support even when direct treatment is unavailable), 2) to the family (to inform subsequent reproductive decisions), and 3) to society (to provide knowledge of the true range and incidence of the condition).14

Meanwhile, in recent years some prominent figures in the world of newborn screening—including the director of the National Institute of Child Health and Human Development (NICHD)—have forcefully criticized the principle that “it is appropriate to screen only for conditions for which effective treatment already exists” as a “dogma” that ought to be discarded. They favor a significantly more expansive approach to newborn screening, in which all conditions—no matter how rare, poorly understood, or currently untreatable—are presumed to be eligible for screening unless specifically excluded on a case-by-case basis.15 At the same time, the NICHD is funding efforts to move beyond today's limited, phenotypic methods of newborn screening toward DNA-based platforms that can “offer enormous opportunities to identify staggering numbers of potentially pathogenic mutations in a very large number of disease-associated genes.”16 Clearly, proponents of this change understand that, if the principle “screen only if you can effectively treat” is set aside and if the technology of newborn screening shifts to primarily DNA-based multiplex platforms, such as gene chips or even whole genome sequencing, the stage will be set for a vast expansion in newborn screening. The new principle guiding newborn screening would then be “screen unless there is a compelling reason not to screen.”

II. Newborn Screening and the ACMG's Expanded Uniform Panel

This scenario is no longer entirely hypothetical; change has already come to the practice of newborn screening in this country. In the past few years, newborn screening has undergone rapid expansion throughout the United States in accordance with recommendations made by the ACMG in its 2005 report, Newborn Screening: Toward a Uniform Screening Panel and System.17 That document is the final report of an ACMG working group commissioned and funded in 2002 by the federal Maternal and Child Health Bureau (MCHB) of the Health Resources and Services Administration, a division of HHS. The ACMG's task was to gather evidence on the effectiveness of newborn screening, to recommend a uniform panel of conditions that ought to be screened for in every state, and to consider other critical components of the newborn screening system.18 In its report, the ACMG recommended that all state-based newborn screening programs adopt a uniform panel of twenty-nine core conditions as well as twenty-five secondary conditions.19 Their recommendation was promptly endorsed by the Advisory Committee on Heritable Disorders and Genetic Diseases in Newborns and Children in a June 2005 letter to Michael O. Levitt, Secretary of HHS.20 By November 2008, almost all of the states had adopted the ACMG's panel of twenty-nine core conditions,21 and most had initiated screening for a majority of the twenty-five secondary conditions.22 As these numbers indicate, the states have moved with unprecedented speed to implement a newborn screening system that is both considerably more uniform and considerably expanded compared to even a few years ago. For comparison, as recently as 2005 (the year the ACMG report was released), the states varied widely in their use of newborn screening tests, “with some mandating screening for as few as three conditions and others mandating as many as forty-three conditions.”23

In light of these extraordinarily rapid developments, the question now before us is, what, if any, alterations in the ethical principles of newborn screening have occurred in the course of the expansion recommended by the ACMG and implemented by most of the fifty states? In particular, is that expansion consistent with the classical principles that have governed newborn screening for the past forty years? Or has there been a break with those principles, and, if so, how significant a break? Answering these questions requires a close look at some features of the ACMG's complex and lengthy report.24

In carrying out its mandate, the ACMG working group evaluated eighty-four heritable disorders for possible inclusion in a uniform newborn screening program that all fifty states would be encouraged to adopt. After conducting a broad survey of expert opinion, the working group assigned a numerical score to each condition, with a high score indicating that the condition was a plausible candidate for mandatory screening. The eighty-four conditions were initially divided into three groups, composed of high-, middle-, and low-scoring conditions. In a second tier of analysis, each condition's initial ranking was re-evaluated by a small number of experts, after which the conditions were assigned to one of three final categories: a core panel of conditions meriting mandatory screening (twenty-nine conditions); a secondary panel of conditions not meeting the standards of the core panel but deemed appropriate for screening anyway (twenty-five conditions); and the remaining conditions, deemed not appropriate for screening at this time.25

A. Reactions to the ACMG Report

As we have seen, the ACMG report and its recommendations received strong support from the HHS Secretary's Advisory Committee on Heritable Disorders and Genetic Diseases in Newborns and Children. It also was endorsed by the American Academy of Pediatrics,26 by the March of Dimes,27 and by other advocacy groups and professional organizations. At the same time, the methods by which the ACMG working group arrived at its recommended screening panels have been faulted on a number of grounds, leading some critics to conclude that the expansion of newborn screening is proceeding too rapidly and without sufficient deliberation and caution. Others have defended the proposed expansion against these criticisms. Before giving our own assessment of the ACMG's report, let us briefly summarize the objections that have been raised by others.

Commenting on the ACMG's recommendations, pediatrician-ethicist Jeffrey Botkin and colleagues urge a cautious approach to expansion.28 They note that, even in its most celebrated and paradigmatic successes (e.g., PKU), newborn screening has proved to be a mixed blessing, with adverse consequences as well as benefits. That is, although they consider PKU screening to be a clear success, they point to avoidable missteps in its implementation and to some continuing adverse consequences, largely from false positive screening results. They also caution that each genetic illness is unique; that population-wide screening of asymptomatic individuals for uncommon diseases has rarely proved effective; that the benefits and risks must be carefully weighed on a condition-by-condition basis; and that rapid expansion of the uniform screening panel without adequate empirical studies would be unwise.

Responding to the ACMG working group's expanded panel of fifty-four conditions and to the prospect of further expansions as new test modalities become available, Botkin and colleagues strongly urged the merits of implementing newborn screening within a circumscribed research paradigm, involving thorough empirical studies to determine for each disorder whether it is suitable for routine screening. The questions that would need study include: Do the benefits of screening for this disorder outweigh the harms, if any? What are the actual medical, psychological, and social outcomes for infants testing positive for the disorder? How common are false-positive results, and what are their consequences? What are the secondary benefits of screening to the family and to the public, and are they substantial enough to justify screening when the traditional standard of direct medical benefit to the child cannot be met? Such research would be conducted in carefully controlled pilot studies, with the aim of gathering vital information about the risks and benefits involved, well before the implementation of population-wide newborn screening.29

Pediatrician Virginia Moyer and colleagues (including Jeffrey Botkin, among others), writing on behalf of the United States Preventive Services Task Force (USPSTF), have criticized the ACMG working group for failing to “conform to contemporary standards of evidence-based decision-making.”30 They complain that a “technological imperative”31 has driven the ACMG to include in the recommended panel diseases that are poorly understood, untreatable, or both, merely because the technology exists to detect them. Believing that the goal of screening should be to improve the health of affected newborns, they find it ethically questionable to mandate screening “in order to recruit research subjects.”32 They invoke the “time-honored tenet of medicine that clinicians should not order a test if the results will not change clinical management,”33 and they find that many of the conditions the ACMG urges states to screen infants for should be excluded on this basis. They argue that the process by which the ACMG working group evaluated conditions for inclusion was flawed, insofar as it relied on unsystematic reviews and colloquial evidence and made use of dubious or obscure criteria. Applying USPSTF decision-making standards, Moyers and colleagues find that for only a handful of the twenty-nine recommended conditions is there adequate evidence that the benefits of screening would outweigh the harms. They suggest that “state and federal policymakers should further evaluate each condition proposed for screening before recommending that it be included in a mandated screening panel.”34 Finally, Moyers and colleagues urge states that implement the expanded panels to “commit to collecting longitudinal data on infants who test positive,”35 to help us implement in the future truly effective evidence-based screening programs.

Mary Ann Baily and Thomas Murray of the Hastings Center have offered a somewhat different critique of the ACMG's recommended expansion of newborn screening. 36 They emphasize that a responsible newborn screening policy must take into account the opportunity costs of expanded screening, which will inevitably draw scarce resources away from other worthy public health programs and needs. They point out that the true costs of a newborn screening program include not only the price of the test itself (which might seem quite small) but also the cost of “parental education, follow-up of all positives to a definitive diagnosis, treatment of affected children, and ongoing data collection and evaluation.”37 In light of these concerns about allocation of health care resources, Baily and Murray criticize the ACMG for including in their panel “conditions that do not urgently need treatment in the newborn period, or for which no proven treatment is available, or for which the benefit of treatment is much less significant and certain than the benefit of treatment for a condition like PKU.”38 In response to the argument that expanded screening and early diagnosis of obscure conditions will help families of affected infants to “avoid a diagnostic odyssey,” they urge that these benefits be weighed against “the burdens of different kinds of odysseys.” First, there is “the period of anxious searching and wandering” that many families of healthy children will experience between an initial false positive result and the confirmatory testing that eventually reassures them that their children are well. Second, there will be families whose children are diagnosed with a serious genetic disorder and yet never become symptomatic:

Perhaps the child has a mild or subclinical form that was unknown before newborns were routinely screened for the disorder… Meanwhile, the family reorganizes its life around medical monitoring and planning for something terrible that never happens.39

Third, there will be children who are accurately diagnosed with disorders for which there are as yet no proven treatments. Their families may begin a “treatment odyssey, searching the Internet, visiting specialists, running up debt, medicalizing the child's life—only to have that life end in early death anyway.”40 Baily and Murray fear that, with expanded screening, such unhappy medical odysseys will become more widespread.

Rebutting such criticisms, R. Rodney Howell, pediatrician and member of the ACMG's Newborn Screening Steering Committee that supervised the preparation of the 2005 report, observed that controversy over the current expansion is reminiscent of early opposition to PKU screening when it was introduced over forty years ago; but that thanks to that program we now have “a whole generation of young adults with treated PKU who have normal intelligence and are productive members of society.”41 He also pointed to the enormous benefits we have reaped from screening for CH and BIOT. Responding to the criticisms of Botkin and colleagues, Howell acknowledged that an expansion of screening will require a complex infrastructure (to support testing, counseling, education, treatment, and follow-up) that is not yet in place. He defended the efficacy of newborn screening in general, however, and pointed out that “there is little advantage at this time to discuss whether there should be expansion of newborn screening; it is occurring briskly at this very moment.”42 Indeed, it is notable that both those urging caution and those enthusiastically embracing the expansion of newborn screening are more or less in agreement that rapid expansion is already taking place, and even more accelerated expansion in the future is all but inevitable.

Howell acknowledged the lack of controlled trials for the treatment of some of the serious metabolic disorders included in the recommended panel, but he insisted that, when the evidence is clear that untreated infants face grave illness or death and that treatment has some efficacy, “no prudent physician would fail to provide treatment information to such families if the condition had been identified.”43 Addressing the issue of false positives, Howell acknowledged that they are problematic and called for more research to keep their numbers to a minimum. Regarding informed parental consent, Howell considered it desirable in the case of poorly understood conditions, but he noted that it is a daunting challenge to try to explain to parents the array of tests to be performed and the various potential outcomes of accepting or rejecting them. Finally, Howell predicted that, with new technologies and new treatments, the number of genetic disorders for which newborns will be screened in the future will expand far beyond the currently recommended panel, and he urged the nation to get to work building the infrastructure that will support such programs, which will have enormous potential benefits for infants in the future.

The issues involved in this debate are complex, and there are numerous cogent arguments on both sides of the controversy. Here, however, we shall focus our attention on two distinctive features of the ACMG's approach: their use of a broadened conception of benefit to justify newborn screening, and their readiness to allow progress in multiplex screening technology to dictate the pace and scope of the expansion of newborn screening. Before turning to these issues, let us first examine the principles that guided the composition of the ACMG's core and secondary screening panels.

B. The Core or Primary Conditions

On a first reading, the ACMG report conveys a strong impression of having followed accepted screening principles in recommending its expanded uniform panel. Consider the following clear statement of a “basic principle developed at the onset of the project”:

To be included as a primary target condition in a newborn screening program, a condition should meet the following minimum criteria:

• It can be identified at a period of time (24 to 48 hours after birth) at which it would not ordinarily be clinically detected.
• A test with appropriate sensitivity and specificity is available.
• There are demonstrated benefits of early detection, timely intervention, and efficacious treatment.44

Elsewhere the report affirms that, when evaluating a disorder for inclusion in the screening panel, “benefit to the child being screened is the overriding consideration.”45 Ultimately, the twenty-nine conditions included in the core panel were those that, in the judgment of the ACMG, met three final criteria. All of them have

1. Specific and sensitive screening tests;

2. A sufficiently well understood natural history; and

3. Available and efficacious treatments. 46

And indeed, a review of the twenty-nine conditions designated “core” or “primary” reveals that, in every case, the ACMG working group concluded that effective treatment was available that could prevent all (for four conditions), most (ten), or, at any rate, some (fifteen) of the disease's symptoms; they also determined that there was clear (fourteen) or some (fifteen) evidence that treatment would benefit the affected newborn. Finally, for twenty-five of the twenty-nine core conditions, the ACMG concluded that the available treatment was efficacious at preventing mortality, independent of any reduction in morbidity.47

Whether all twenty-nine of the core conditions do actually meet the chief criterion—the availability of an effective treatment that will clearly benefit those newborns who test positive for the condition—is open to debate.48 In this context, it is significant that a comprehensive study of the effectiveness and cost-effectiveness of newborn screening for metabolic disorders in the United Kingdom concluded in 2004 that the evidence supports screening for only two of the conditions detectable by MS/MS: PKU and MCAD. As for the other conditions identifiable by MS/MS, the authors concluded the following:

Robust evidence on the underlying incidence and outcomes for many of the disorders was lacking, particularly differences in long-term outcomes that could be attributed to therapies initiated as a consequence of presymptomatic detection using tandem MS.49

The British recommendation to screen only for two metabolic disorders was based on a systematic analysis of the same evidence that led the ACMG working group, a year later, to include twenty conditions detectable by MS/MS in its core panel of twenty-nine, and twenty-two more in its secondary panel of twenty-five conditions—to which we now turn our attention.

C. The Secondary Conditions

Besides the twenty-nine core or primary conditions, the ACMG report also recommended mandatory screening for twenty-five “secondary” conditions that did not meet the three criteria (listed in the preceding section) for inclusion among the core conditions. More precisely, these are conditions that, despite the availability of a specific and sensitive screening test, lacked a well understood natural history, an efficacious treatment, or both. Why, then, were they recommended for mandatory screening?

When first introduced in the ACMG report, the secondary conditions are described as “conditions that are part of the differential diagnosis of a core panel condition.”50 In other words, some core conditions, in order to be diagnosed with precision, require the gathering of data that also would reveal the presence in the newborn of one or more other conditions—conditions whose natural history is poorly understood or for which effective treatment is not currently available. The necessity of screening for these secondary conditions would therefore seem to be a mere accident of the testing protocol, an unintended and even regrettable consequence of the determination to screen for the core conditions. The ACMG working group concluded that positive results for these secondary conditions ought to be reported by the laboratory to the health care provider and to the family of the infant, presumably on the grounds that clinically significant results, once obtained, could not ethically be withheld from the newborn's physician and parents.51

It may indeed be ethically problematic to withhold clinically significant test results for the secondary conditions, even if they were included in the screening panel only because they show up as part of the differential diagnosis of core conditions. For most of these conditions, however, the clinical significance of a positive screening result is very much in doubt. And it is necessary to consider the consequences of informing physicians and parents that a child has been identified as having a serious genetic disorder, when the natural history of that disorder is poorly understood and the appropriate treatment—or even the need for treatment—is highly uncertain.

An instructive case is Wisconsin's experience with the organic acid disorder 2-methylbutyryl-coenzyme A dehydrogenase deficiency (2-MBG), now included in the ACMG's secondary panel because it is part of the differential diagnosis (by MS/MS) of the core condition IVA.52 As of November 2008, forty-three states were screening newborns for 2-MBG. Outside of the United States, only a handful of infants have been diagnosed with 2-MBG, some of whom have suffered severe developmental delay, failure to thrive, seizures, muscle atrophy, and/or cerebral palsy, while others have remained entirely asymptomatic. Some of those infants have been treated with a restricted diet, with inconclusive results. Only five cases of 2-MBG had been described worldwide when the state of Wisconsin, in 2000, added 2-MBG to its mandatory newborn screening panel. Surprisingly, over the next six years, twenty-seven Wisconsin infants were identified with 2-MBG, all but one of whom were offspring of Hmong parents.53 Most of these infants were put on a diet low in protein and supplemented with carnitine, although parental compliance with this diet was quite variable. In any event, as of 2007, all of the Wisconsin children who had been diagnosed at birth with 2-MBG were normal, healthy, and asymptomatic. It is at this point uncertain whether the restrictive diet was effective, whether the genotypic variant of 2-MBG shared by Hmong infants is essentially benign, and whether infants identified at birth with 2-MBG ought to be treated presymptomatically at all. What is clear is that nationwide screening for 2-MBG, as recommended by the ACMG, will result in a significant number of children (mostly of Hmong descent) being labeled with a serious illness, despite the fact that the majority of them might remain asymptomatic even without any treatment.

We cannot say how typical the case of 2-MBG is of the conditions on the ACMG's secondary screening panel. However, with twenty-five of these rare and poorly understood disorders being proposed to the states for mandatory screening, and with more presumably on the way as the panel of primary conditions expands further, the number of American children in this doubtful situation seems destined to grow considerably. In addition to the possibility of over-treating (and possibly harming) children who are healthy, there also is concern about contributing to “vulnerable child syndrome” and parental overprotectiveness.54 Thus, the call for a mandatory secondary screening panel—however necessary it may be for differential diagnosis of the core conditions—is a proposal fraught with unintended but possibly serious consequences.

So far, at least, it would appear that the recommendation to screen for a panel of secondary conditions—poorly understood, not clearly treatable, or both—is merely an artifact of the way some primary conditions are detected, an unfortunate necessity that could be avoided altogether if there were a way to screen only for the primary condition. Only twenty of the twenty-five secondary conditions, however, were included in the panel on this basis. When spelled out fully, the ACMG's standard for including a condition in the secondary panel is that the condition must be either “part of the differential diagnosis of a primary target condition” or merely “apparent in the result of the multiplex assay.”55 In fact, five conditions were evidently added to the secondary panel on this less stringent basis.56 Each of these five conditions is very far from meeting the classical standards for inclusion in a newborn screening program.

For example, one of the five is the fatty acid oxidation defect, dienoyl-CoA reductase deficiency (DE-RED). The incidence of this disorder is unknown, but it must be extremely rare, as only one case of it has ever been reported. It therefore is impossible to be certain whether the symptoms exhibited by that one infant were coincidental or in fact were caused by the genetic defect. The ACMG report comments that “the sensitivity and specificity of the primary marker are also unknown,”57 as are the availability, cost, and potential efficacy of any treatment. DE-RED is not detected as part of the differential diagnosis of any other condition; nonetheless, simply because it is possible to detect DE-RED using MS/MS, the ACMG report recommends that DE-RED screening be included in the mandatory (secondary) screening panel in all fifty states. As of November 2008, newborn screening for DE-RED was mandated by law in sixteen states and offered in three others.

D. The Role of Multiplex Screening Platforms

To understand why these five rare, poorly understood conditions were included in the ACMG's secondary target category, it is necessary to delve more deeply into the technology of present-day newborn screening. Some conditions that are candidates for newborn screening are identified by way of unique testing methodologies, but many can be detected using multiplex platforms that screen simultaneously for several conditions. Of the twenty-nine core conditions, twenty-three are identified using multiplex platforms: MS/MS for the six amino acid disorders, the nine organic acid disorders, and the five fatty acid oxidation disorders; and either high pressure liquid chromatography (HPLC) or isoelectric focusing (IEF) for the three hemoglobinopathies. Only six of the core conditions require unique “singleton” tests.58 Of the twenty-five secondary conditions, all but two are detectable using multiplex platforms.59

The ACMG report emphasizes the advantages of multiplex screening technology:

Particularly notable is the implementation of multiplex platforms that allow a single type of specimen preparation and simultaneous (or nearly simultaneous) screening for multiple different disorders. Going from one test for one disorder to one test for multiple disorders has the potential to reduce costs per condition tested and can lead to test expansion if these new technologies can be integrated safely and effectively into newborn screening programs.60

The report notes that, with some multiplex platforms, the screener must select specific targets for inclusion in the test (this is known as “selective reaction monitoring” or SRM), while for others the test automatically screens for multiple targets without the need for specific target selection (this is known as “full profile testing”).61 MS/MS in particular can be used in either selective or full profile mode.62 Selective monitoring means using the multiplex platform to target only those conditions deemed appropriate for screening. In contrast, the full profile approach means making maximum use of the technology's information-gathering powers, without regard to the distinction between appropriate and inappropriate target conditions.

Remarkably, the ACMG report makes a forceful case that, whenever possible, MS/MS screening should be carried out in full profile mode. The report gives several reasons for this judgment, one of which is simply that “the use of MS/MS profiles allows for the maximal use of the technology for the identification of clinically significant conditions.”63 Elsewhere the report extols “the inherent value of multiplex technologies to public health.”64 But why should “maximal use of the technology for the identification of clinically significant conditions” be considered inherently good, when some of the conditions that will be detected are considered inappropriate for screening? Here is the report's answer to that question:

Although information about conditions for which treatment options are scarce or not yet reported can lead to increased stresses on families and the health care system, early information can also lead to knowledge of the condition for the family, thus avoiding a potential diagnostic odyssey or inappropriate therapies. In addition, early information provides opportunity for better understanding of disease history and characteristics, and for earlier medical interventions that might be systematically studied to determine the risks and benefits. Multiplex testing and the identification of conditions falling outside of the uniform screening panel provides the opportunity for such conditions to be included in research protocols.65

In other words, screening for a condition that fails to meet the classical criteria can be justified under a broadened conception of benefit that includes not only helping the family avoid “the diagnostic odyssey” but also helping society by providing opportunities for biomedical research aimed at understanding the natural history of the disorder and finding an effective treatment for it.

E. A Broadened Conception of Benefit

The ACMG's emphatic preference for the use of multiplex platforms in “full profile” mode is thus indicative of a broadened conception of benefit that could justify screening for nearly any condition. Traditionally, as we have seen, the only relevant benefit was the benefit to the infant of a timely and effective treatment for a serious illness. The ACMG report, on the other hand, is quite explicit in embracing a broader notion of public benefit, not limited to direct treatment of the child. In assessing each testable condition for inclusion in the uniform panel, the authors of the ACMG report gave “overriding consideration” to benefits of early intervention for the individual screened (chiefly when there is a known and effective treatment), but they also gave weight to “benefits of early intervention for family and society”:

Families could benefit from establishing that there may be a genetic risk to others in the family. Society could benefit by a reduction in medical diagnostic odysseys that are costly to the healthcare system.66

Elsewhere, the report makes clear that the societal benefits of newborn screening include the opportunity for progress in biomedical research.67 Thus, it seems clear that, in extolling the advantages of multiplex platforms and in calling for their use in “full profile” mode—even when some of the conditions detected are rare, poorly understood, and as yet untreatable—the ACMG working group was thinking more broadly about benefits to family and to society, and especially about the value of studying rare and obscure disorders in order to understand them and to find an effective treatment.

F. Impact of a Broadened Conception of Benefit on the ACMG Recommendations

The impact of this broadened conception of benefit on the ACMG working group's final recommendations can be seen in two different places: in the initial scoring of the conditions that assigned them to one of three categories (low-, middle-, and high-scoring), and in the way certain conditions found their way into the secondary panel.

When surveying expert opinion for the initial ranking of the eighty-four conditions that were candidates for screening, the ACMG asked a series of questions and assigned points to each condition based on the answers of hundreds of experts. With extremely favorable answers from the experts, a condition could score a maximum of 2,100 points. Of these, up to 700 points were awarded to a condition based on some of the main attributes of the condition (incidence, burden if untreated, benefits of intervention, etc.), up to 700 points for attributes of the screening test (sensitivity, specificity, multiplex versus singleton, etc.), and up to 700 points for aspects of treatment and management (availability, cost, efficacy, etc.). Among many other criteria, up to 200 points were awarded to a condition for evidence of “individual benefits of early intervention.” But up to 100 points were awarded to a condition for evidence of “family and societal benefits of early intervention.” Moreover, another 200 points were awarded to a condition just for being detectable using a multiplex platform, and fifty additional points were awarded to a condition if, in the course of detecting it, other conditions were also identified. In other words, a condition might well be bumped up to the middle- or even to the high-scoring division despite showing scant evidence of benefit to the newborn child.

After the initial scoring, the eighty-four conditions were re-evaluated using a decision tree that is depicted in Figure 9 of the ACMG report (reproduced below). High-scoring (>1,200) conditions were added to the core panel if, on further review, experts determined that a treatment was available and necessary and that the natural history of the disease was well understood. But high-scoring conditions for which there was no treatment still ended up in the secondary panel if they were part of the differential diagnosis of a core condition, or even if they were merely detectable as part of a multiplex assay in full profile mode. If a treatment was available but the natural history of the disease was poorly understood, the high-scoring condition still ended up in the secondary panel. And indeed, the report notes that three conditions that initially scored high on the survey “were moved to the secondary target category on the basis of scientific evidence indicating that the natural history was not sufficiently well understood.”68 In other words, the lack of an effective treatment or of an adequate understanding of the natural history of a disease (or both) was not sufficient to remove it from the mandatory screening panel; it merely led to the disease being re-classified as a secondary rather than a primary target.

Meanwhile, middle-scoring conditions (1,000-1,200) were added to the secondary panel as long as they were part of a differential diagnosis or were detectable by multiplex assay. And even low-scoring conditions (<1,000) were to be bumped up to the secondary panel if they were detectable in a multiplex assay.69 In this regard, it is worth noting that the ACMG working group considered for inclusion but ultimately rejected the rare lysosomal storage disorders Fabry, Krabbe, Pompe, and Hurler-Scheie, scores of which ranged from 447 to 707. These extremely low scores reflect not only the absence, at present, of a sensitive and specific screening test, but also the unavailability of any effective treatments. Nevertheless, researchers are well on their way to developing MS/MS assays for these conditions,70 and the ACMG report's decision tree would seem to dictate the automatic inclusion of all these untreatable and poorly understood disorders (and related disorders such as Gaucher, Hunter, Niemann-Pick, and Tay-Sachs) in the secondary panel as soon as it becomes feasible to detect them using a multiplex platform such as MS/MS.

Figure 9 from the ACMG's Newborn Screening Report71

Copyright© American College of Medical Genetics 2006. Unauthorized reproduction is prohibited.

III. The Significance of the ACMG's Recommended Screening Panel

In considering this decision-making process, it is important to bear in mind that whether a state program mandates screening for a condition as part of the core panel or part of the secondary panel makes no practical difference as far as the infant and family are concerned. If a state screening program embraces the ACMG's recommendation of mandatory screening for both primary and secondary conditions (and most states seem to be on their way to doing so), a positive screening result is reported to the physician and the family regardless of the target category in which the condition happened to be included. With this in mind, the ACMG report's decision tree seems to depart much more radically from classical screening principles than it first appeared to do.

As noted above, the twenty-nine core conditions were each judged by the ACMG working group to meet the traditional standard of having 1) a specific and sensitive screening test, 2) a sufficiently well understood natural history, and 3) an available and efficacious treatment. Moreover, for twenty of the twenty-five secondary conditions, screening could be considered justified on the grounds that it was necessary for the differential diagnosis of one of the core conditions. Only five exceedingly rare conditions were added to the secondary panel without that compelling justification. Seen in this light, the expansion in newborn screening recommended by the ACMG would appear to be rather moderate and fairly consonant with accepted screening principles.

But a careful examination of the ACMG working group's decision tree—and, above all, of its procedures for adding conditions to the secondary panel—makes it clear that the foundations have been laid for a much more radical expansion of newborn screening in the future, and for a significant loosening of the traditional screening standards. Under the ACMG's procedures, a rare and poorly understood genetic condition, even one with no available treatment, will routinely be added to the secondary target panel (recommended for mandatory screening in all fifty states) as soon as it becomes possible to detect that disorder using a multiplex assay in full profile mode. Even if only a handful of conditions have so far qualified for the secondary panel under that rubric, it is clear that many more conditions could be added to the panel in the future, especially if rapid progress is made in the exploitation of DNA-based multiplex screening platforms, with their potential to detect hundreds of thousands of genetic abnormalities at one stroke.

In brief, it seems fair to conclude from a careful reading of the report that the ACMG working group has effectively recommended mandatory newborn screening for two categories of conditions: the relatively small number of treatable and well understood disorders that satisfy the classical Wilson-Jungner criteria, and the potentially much larger set of untreatable and poorly understood disorders that fall short of those criteria but are detectable by multiplex screening. If the chief purpose of screening conditions in the former category is to benefit the affected newborn with timely and effective treatment, the chief purpose of screening conditions in the latter category would seem to be to advance the scientific study of the disorder, with the ultimate goal of finding an effective treatment. That is certainly a laudable goal, but as a basis for including conditions in a mandatory newborn screening panel it represents a sharp departure from the principles expressed in the Wilson-Jungner criteria. Hitherto, for diseases that were poorly understood or for which no effective treatment was available, we as a nation have not been in the habit of subjecting individuals to compulsory screening merely for research purposes. In the wake of the ACMG report and its enthusiastic reception by the states, our approach to newborn screening seems to be heading into uncharted territory.

_______________________

Endnotes

1. James M. G. Wilson and Gunnar Jungner, Principles and Practice of Screening for Disease (Geneva: World Health Organization, 1968), available online at whqlibdoc.who.int/php/WHO_PHP_34.pdf.

2. By case-finding, Wilson and Jungner mean “that form of screening of which the main object is to detect disease and bring patients to treatment, in contrast to epidemiological surveys.” In contrast, the main purpose of surveys is “not to bring patients to treatme nt but to elucidate the prevalence, incidence, and natural history” of the disease or symptom under study. (Ibid., p. 12.)

3. Ibid., pp. 26-27.

4. Ibid., p. 27; emphasis added.

5. Ibid., pp. 27-28; emphasis added.

6. Ibid., p. 32.

7. Ibid., p. 28.

8. William K. Frankenburg, “Selection of Diseases and Tests in Pediatric Screening,” Pediatrics 54 (1974): 612-616.

9. Ibid., p. 616.

10.Committee on Assessing Genetic Risks, Institute of Medicine, Assessing Genetic Risks: Implications for Health and Social Policy, Lori B. Andrews, Jane E. Fullarton, Neil A. Holtzman, and Arno G. Motulsky, eds. (Washington, D.C.: National Academies Press, 1994).

11. American Society of Human Genetics and American College of Medical Genetics, “Points to Consider: Ethical, Legal, and Psychosocial Implications of Genetic Testing in Children and Adolescents,” American Journal of Human Genetics 57 (1995): 1233-1241, p. 1233.

12. Neil A. Holtzman and Michael S. Watson, Promoting Safe and Effective Genetic Testing in the United States: Final Report of the Task Force on Genetic Testing (Bethesda, Maryland: National Institutes of Health, 1997), available online at http://biotech.law.lsu.edu/research/fed/tfgt/.

13.American Academy of Pediatrics Newborn Screening Task Force, “Serving the Family from Birth to the Medical Home. Newborn Screening: A Blueprint for the Future,” p. 394.

14.National Research Council Committee for the Study of Inborn Errors of Metabolism, Genetic Screening Programs, Principles, and Research (Washington, D.C.: National Academy of Sciences, 1975), p. 1.

15. Duane Alexander and Peter C. van Dyck, “A Vision of the Future of Newborn Screening,” Pediatrics 117 Supplement (2006): S350-S354.

16. Duane Alexander and James W. Hanson, “NICHD Research Initiative in Newborn Screening,” Mental Retardation and Developmental Disabilities Research Reviews 12 (2006): 301-304, p. 302.

17. American College of Medical Genetics, Newborn Screening: Toward a Uniform Screening Panel and System ( Washington, D.C. : Health Resources and Services Administration, 2005), hereafter cited as ACMG, Newborn Screening. The full report is available online at http://mchb.hrsa.gov/screening (the version cited here) and (in a version published as a supplement to Genetics in Medicine 8 [2006]: pp. 1S-252S) at www.acmg.net/resources/policies/NBS/NBS-sections.htm.

18. ACMG, Newborn Screening, p. 7.

19. The crucial distinction between core and secondary conditions is explained later in the chapter. For the present, let the primary or core conditions be understood as those that fully meet the criteria for inclusion in the uniform screening panel, while the secondary conditions are those that fall short of that standard but—according to the ACMG—merit inclusion in the panel on other grounds.

20. In its letter, the Advisory Committee “strongly and unanimously recommends that the Secretary initiate appropriate action to facilitate adoption of the ACMG recommended screening panel by every State newborn screening program.” The letter may be found online at www.hrsa.gov/heritabledisorderscommittee/reports/letterstoSecretaryofHHS.htm.

21. More precisely, as of November 2008, all of the states screen for at least twenty-six of the twenty-nine core conditions, and forty-four states screen for all of them. See the screening statistics compiled by the National Newborn Screening and Genetics Resource Center (NNSGRC) at http://genes-r-us.uthscsa.edu/nbsdisorders.pdf. Note that the NNSGRC status report may not accurately reflect the mandatory or voluntary status of newborn screening in each state. For example, Massachusetts currently offers two screening panels, a mandatory panel of ten conditions and an optional panel of twenty others.

22. More precisely, as of November 2008, thirty states are now or soon will be screening for more than twenty of the twenty-five secondary conditions; at the other end of the spectrum, three states (Arkansas, Kansas, and Louisiana) screen for only two of the secondary conditions.

23. ACMG, Newborn Screening, p. 7. According to a November 2008 report by the CDC, “After 2006, most states began to expand their panels to include all 29 disorders; currently, 21 states and the District of Columbia have fully implemented the ACMG panel.” The CDC also analyzed newborn screening data from 2001 to 2006 from states with well-established MS/MS screening programs to “estimate the number of children in the United States who would have been identified with disorders in 2006 if all 50 states and the District of Columbia had been using the ACMG panel.” This analysis led the CDC to conclude that such an expansion would have increased the number of children identified in 2006 by only thirty-two percent (from 4,370 to 6,439). But the additional children identified would have had “many rare disorders that require local or regional capacity to deliver expertise in screening, diagnosis, and management.” According to the CDC's analysis, nine of the disorders detectable by MS/MS each accounted for an estimated fifteen cases or fewer; in contrast, SCD and CH (both already targeted by population-wide screening in the United States) together accounted for sixty-one percent of the total estimated number of cases. See Centers for Disease Control and Prevention, “Impact of Expanded Newborn Screening—United States, 2006,” Journal of the American Medical Association 300 (2008): 2242-2244.

24. The report itself is 108 pages long, with an additional 221 pages of figures and appendices.

25. ACMG, Newborn Screening, pp. 9-10. The working group deferred a decision regarding screening for three infectious diseases included among the eighty-four conditions evaluated: human HIV infection, congenital toxoplasmosis, and congenital cytomegalovirus infection. See p. 66.

26. “AAP Endorses Newborn Screening Report from the American College of Medical Genetics,” press release of May 12, 2005, available online at www.aap.org/advocacy/releases/mayscreening.htm.

27. “March of Dimes Statement on Newborn Screening Report,” September 22, 2004, available online at www.marchofdimes.com/aboutus/10651_13507.asp.

28. Jeffrey R. Botkin, et al., “Newborn Screening Technology: Proceed With Caution,” Pediatrics 117 (2006): 1793-1799.

29. Of course, for the rarer conditions, affecting fewer than one in 10,000 newborns, it will prove quite difficult to conduct statistically valid research on smaller subpopulations prior to full-scale screening. For a proposal for a structured sequence of research protocols to evaluate potential applications for newborn screening before their formal implementation in public health programs, see Jeffey R. Botkin, “Research for Newborn Screening: Developing a National Framework,” Pediatrics 116 (2005): 862-871.

30. Virginia A. Moyer, et al., “Expanding Newborn Screening: Process, Policy, and Priorities,” Hastings Center Report 38 (2008): 32-39, p. 33.

31. Ibid., p. 33.

32. Ibid., p. 34.

33. Ibid., p. 35.

34. Ibid., p. 33.

35. Ibid., p. 39.

36. Mary Ann Baily and Thomas H. Murray, “Ethics, Evidence, and Cost in Newborn Screening,” Hastings Center Report 38 (2008): 23-31.

37. Ibid., p. 27.

38. Ibid., p. 28.

39. Ibid., p. 29.

40.Ibid., p. 29.

41. R. Rodney Howell, “We Need Expanded Newborn Screening,” Pediatrics 117 (2006): 1800-1805, p. 1802.

42. Ibid., p. 1802.

43. Ibid., p. 1802.

44. ACMG, Newborn Screening, p. 28.

45. Ibid., p. 42.

46. Ibid., p. 62. Note that these three benchmarks correspond roughly to the fifth, seventh, and second Wilson-Jungner criteria, respectively.

47. These numbers are collected in a helpful review of the ACMG report by Donald Bailey and colleagues; Donald B. Bailey, Jr., et al., “Changing Perspectives on the Benefits of Newborn Screening,” Mental Retardation and Developmental Disabilities Research Reviews 12 (2006): 270-279, p. 273, Table 1.

48. In addition to the article by Moyer, et al., “Expanding Newborn Screening: Process, Policy, and Priorities,” see Marvin Natowicz, “Newborn Screening—Setting Evidence-Based Policy for Protection,” New England Journal of Medicine 353 (2005): 867-870.

49. Abdullah Pandor, et al., “Clinical Effectiveness and Cost-Effectiveness of Neonatal Screening for Inborn Errors of Metabolism Using Tandem Mass Spectrometry: A Systematic Review,” Health Technology Assessment 8 (2004): 1-121.

50. ACMG, Newborn Screening, p. 8.

51. Ibid., pp. 11, 20, 62, 75. Rodney Howell (“We Need Expanded Newborn Screening,” p. 1801) says that the ACMG's experts felt strongly that any serious abnormality revealed by the secondary screening panel “should not be kept secret” from the child's physician and parents. He contrasted this approach with that of the German screening program, “in which information about conditions not listed on their panel would not only be withheld but that the information should be destroyed.” According to Rodney J. Pollitt, in 2004 the German Federal Ministry for Health and Social Security approved screening newborns for fourteen disorders, but the government also forbade screening for any other condition and “decreed that accidentally obtained results arising from the allowed screens must be ignored and not communicated to anyone. All blood-spot samples are to be destroyed within three months.” Rodney J. Pollitt, “International Perspectives on Newborn Screening,” Journal of Inherited Metabolic Disease 29 (2006): 390-396, p. 392.

52. Sandra C. van Calcar, et al., “2-Methylbutyryl-CoA Dehydrogenase Deficiency in Hmong Infants Identified by Expanded Newborn Screen,” Wisconsin Medical Journal 106 (2007): 12-15. We are grateful to Norma n Fost for bringing this case to our attention.

53. Van Calcar, et al., estimate that, in Wisconsin, 2-MBG has a prevalence of 1:223 among Hmong infants, but only 1:325,593 among white newborns. See van Calcar, et al., “2-Methylbutyryl-CoA Dehydrogenase Deficiency in Hmong Infants Identified by Expanded Newborn Screen.” As of 2008, eighty-one cases of 2-MBG had been identified in the United States, seventy-two of them in Wisconsin and Minnesota, where large Hmong populations have settled. (Data downloaded from the National Newborn Screening Information System, available online at www2.uthscsa.edu/nnsis.)

54. On parental attitudes toward the “vulnerable child,” see Morris Green and Albert J. Solnit, “Reactions to the Threatened Loss of a Child: A Vulnerable Child Syndrome,” Pediatrics 34 (1964): 58-66; Morris Green, “Vulnerable Child Syndrome and Its Variants,” Pediatrics in Review 8 (1986): 75–80; and Michael Thomasgard and W. Peter Metz, “Parental Overprotection and Its Relation to Perceived Child Vulnerability,” American Journal of Orthopsychiatry 67 (1997): 330-335. According to Thomasgard and Metz, families of children who recovered from a potentially fatal medical condition were found “to share a number of characteristics that centered on the marked discrepancy between the child's normal health, growth, and development, and the parent's unfounded belief that the child continued to be at risk for serious illness and was destined to die prematurely.”

55. ACMG, Newborn Screening, p. 9.

56. Ibid., p. 64. The report indicates that only four rare conditions were moved into the secondary target panel for this reason (viz., that they are detectable in a multiplex assay, although detecting them is not required for the differential diagnosis of any primary condition): the fatty acid oxidation disorders Short-chain acyl-CoA dehydrogenase deficiency (SCAD) and Dienoyl-CoA reductase deficiency (DE-RED), the organic acid disorder Isobutyryl-CoA dehydrogenase deficiency (IBG), and the amino acid disorder Argininemia (ARG). But it appears that the organic acid disorder Malonic acidemia (MAL) also was included in the secondary panel on this basis.

57. Ibid., p. 233.

58. Ibid., p. 10. The six exceptions are CH, BIOT, CAH, GALT, hearing loss, and CF.

59.Those two are the carbohydrate disorders galactokinase deficiency (GALK) and galactose epimerase deficiency (GALE), conditions included in the secondary panel because they are a part of the differential diagnosis of GALT.

60. Ibid., pp. 18-19.

61. Ibid., p. 19. An example of selective multiplex screening is the use of “gene chips” to test for the presence of a specific set of single nucleotide polymorphisms (SNPs) in the human genome; an example of full profile testing is DNA-sequencing of the entire genome.

62. Ibid., pp. 19, 60-61.

63.Ibid., p. 61. The other reasons are the following: 1) Most of the other conditions identifiable by MS/MS are already required for the differential diagnosis of the high-scoring core conditions; thus full profile screening ensures that the core conditions are diagnosed with the maximum specificity and sensitivity. 2) “Allowing all information to be assessed can reveal the presence of spurious signals and/or contaminants.” 3) Full profile testing can “enhance clinical interpretation of results by revealing anomalies in associated compounds or in compounds that provide internal standards against which excesses or deficiencies can be better interpreted.”

64. Ibid., p. 51.

65. Ibid., p. 20; emphasis added.

66. Ibid., p. 43. On expansive notions of benefit in the ACMG's report, see also Donald B. Bailey, Jr., et al., “Changing Perspectives on the Benefits of Newborn Screening.”

67. Note the phrase “understanding prevalence and natural history” in Table 2 on page 44 of the ACMG's report, as well as the earlier reference, on page 20, to the “opportunity for better understanding of disease history and characteristics, and for earlier medical interventions that might be systematically studied to determine the risks and benefits.”

68. Ibid., p. 64. In fact, all three of these conditions are identified as part of the differential diagnosis of one of the core conditions, so they would have been included in the secondary panel regardless of their survey score.

69. In fact, no condition scoring <1,000 was moved to the secondary panel on this basis; the lowest-scoring condition in the secondary panel is Citrullinemia type II (CIT-II, score 1,001), included because it is in the differential diagnosis of Citrullinemia type I (CIT-I). However, it is to be noted that, of the thirty disorders specifically excluded from screening, fully twenty-three of them were rejected simply because no reliable test is currently available. Many of these conditions would presumably be moved into the secondary panel (and some even to the core panel) if a test became available, especially if it were part of a multiplex assay.

70. See the references in Chapter One, Footnote Thirteen.

71. Ibid., p. 121.

NEXT CHAPTER