Newborn Screening for Metabolic Disorders

Four million newborn infants will receive expanded screening for metabolic disorders in the United States each year; 12 000 will receive a false-positive screening result and more than 1000 will be diagnosed with a metabolic disorder.^{1 - 2} While frequent hospitalizations, developmental delay, mental retardation, and death will be prevented in many of these diagnosed children, many others will have a disorder that is mild or benign. Despite uncertainty, expanded newborn screening for a wide variety of rare disorders appears to be here to stay. Currently, 47 states mandate or offer expanded screening.³ The need for research on these conditions and screening programs has never been more pressing.

In this issue of JAMA, van Maldegem and colleagues⁴ report their findings on short chain acyl-coenzyme A dehydrogenase (SCAD) deficiency occurring in 31 Dutch patients and describe the implications of these findings for newborn screening. These patients were identified because of clinical symptoms ranging from severe and prolonged to mild and transient. Follow-up clinical data were obtained from interviews with physicians or from medical records. Results showed that genotype, while related to biochemical levels of ethylmalonic acid and butyrylcarnitine, did not predict clinical phenotype. The developmental and medical problems that occurred in the children were attributed to other causes in 5 cases. Nine relatives of 10 patients were discovered to have the same mutations as the probands with SCAD deficiency. These siblings and parents were asymptomatic.

van Maldegem et al⁴ noted that SCAD deficiency was not clinically severe and that it was not possible to differentiate diseased from nondiseased individuals. Failure to meet these standard newborn screening criteria⁵ led the authors to question inclusion of SCAD deficiency in expanded newborn screening panels. Since screening for phenylketonuria first began in the 1960s, investigators and policy makers have struggled to justify implementation or discontinuation of newborn screening for particular disorders.⁶ With the current expansion in screening, pediatricians in the community have difficulty making sense of the results reported by screening programs and feel uncomfortable discussing the result with families.⁷

The study by van Maldegem et al⁴ provides useful information on this issue, but their data do not provide enough information to rationalize exclusion of SCAD deficiency from screening panels. First, the lack of association between genotype and clinical phenotype is well known in metabolic disorders. In phenylketonuria, for example, developmental outcomes can differ dramatically in untreated siblings with identical mutations.⁸ Heterogeneity in outcome, despite identical genotypes, also exists in medium chain acyl-coenzyme A dehydrogenase deficiency, with severe consequences including sudden death in some children and no discernible symptoms in others.⁹

Second, descriptions of clinical outcomes based on chart review are often incomplete, if not inaccurate.¹⁰ Symptoms and signs are often presumed nonexistent unless documented in the medical record. The authors used the term “relatively benign course” to describe the outcomes in patients with SCAD deficiency, although more than half the children experienced developmental delay and only 3 were considered to have a normal outcome. Objective measures of intelligence and behavior are needed.

Third, 16 of the 31 children were younger than 6 years and not yet of school age. Conclusions regarding outcome in metabolic disorders require long-term follow-up. For example, early identification and treatment of children with galactosemia reversed or prevented severe complications including liver failure, cataract, coma, and death. The children who were symptomatic were thought to have “recovered completely.” However, long-term follow-up revealed ongoing speech and language deficits, motor problems, and low intelligence quotient in many children.¹¹

And fourth, children identified with SCAD deficiency by newborn screening may represent a different subset of children than those identified because of clinical symptoms, with different genotypes and different outcomes.¹²

Even considering the useful information provided by van Maldegem et al,⁴ 2 issues remain. What criteria should be used to select a condition for expanded newborn screening and what indicators constitute evidence for each criterion? Criteria developed in the 1970s⁵ require updating in light of new technologies, greater understanding of metabolic disorders, and the sociopolitical environment. The rationale that a condition be common no longer seems warranted, given the ease with which disorders can be added when screening is conducted using tandem mass spectrometry.¹³ There is some evidence that the public considers learning disabilities, speech delay, and behavioral difficulties to be serious enough to warrant screening.¹⁴ Moreover, as some researchers, clinicians, and policy analysts have suggested, it is time to rethink the traditional criterion of “treatability.” Newborn screening leads to a reduction in use of diagnostic tests and an increase in knowledge of the genetic basis of the disorder, which in turn may be useful to families in future reproductive decisions.¹⁵ These results are considered justification for screening even when current therapies do not prevent medical or developmental problems.

On the other hand, screening for SCAD deficiency will increase the number of children with false-positive identifications, which are associated with increased parental stress and altered parent-child relationships.¹⁶ However, these consequences appear to be related to the way in which information is communicated to the family. With increased attention to the informing process, this negative consequence of expanded newborn screening may be mitigated or prevented. Nonetheless, the effects of screening and treatment for the conditions justified by new criteria must be evaluated over the long term.¹⁷ Meanwhile, parents clearly value expanded newborn screening. For example, in Massachusetts, where expanded screening is conducted on a pilot basis for SCAD deficiency and other disorders, parents opposed to these tests sign an “opt out” form and these disorders are excluded in the screening. A report from 2001 indicated that fewer than 3% of families opt out.¹⁸

The retrospective study on clinically identified patients with SCAD deficiency by van Maldegem et al⁴ deserves serious consideration. Long-term studies comparing children identified by newborn screening with clinically identified children are needed.¹⁹ In addition to laboratory and genetic tests, long-term follow-up studies should include objective measures of outcome, such as growth, nutritional status, physical well-being, number of hospitalizations, number of emergency department visits, intelligence quotient levels, academic achievement, behavioral characteristics, language development, and motor skills. Treatment and psychosocial variables and factors related to comorbidity should be analyzed. These types of studies require significant funding, interdisciplinary cooperation and collaboration among metabolic centers, newborn screening programs, and public health departments to ensure adequate patient ascertainment and long-term follow-up. Only through comprehensive, long-term research will a rational, fair, and universal newborn screening policy become reality.