Newborn screening in the genomics era: are we ready for genome sequencing?

Recent advances in next generation sequencing (NGS) could potentially revolutionize newborn screening, the largest public health genetics program in the United States and around the world.  Over the last five decades, newborn screening has grown from screening for one condition (phenylketonuria (PKU)) in one state, to nationwide screening for at least 31 severe but treatable conditions, most of which are genetic.  Each year, thousands of babies in the United States are saved from lifelong disability and even death by timely diagnosis and initiation of treatment. An important aspect of newborn screening is speed; many of the diseases that are screened for are inborn errors of metabolism in which the baby’s body cannot properly break down certain substances in food which can build up to toxic amounts.

Most newborn screening tests are biochemical tests that use drops of blood from a heel prick, dried onto a piece of filter paper. The introduction of tandem mass spectrometry made it possible to screen for many conditions at the same time using a single test, and led to a huge expansion in the number of conditions that could be included, with some U.S. states now screening for over 50 conditions.

The initial, first-tier screening tests detect those babies at increased risk of being affected by any of the diseases targeted by the program. Sometimes further screening tests (second-tier tests) are needed to narrow down the group of babies to determine which ones require diagnostic testing. For some conditions, the second-tier test is based on DNA. For example, the first-tier screen for cystic fibrosis looks at concentrations of the immunoreactive trypsinogen (IRT) protein. In some states, genetic screening is then done on the newborn screening blood spots from those babies with high IRT concentrations to look for mutations (changes) in the gene that causes cystic fibrosis. However, this screen currently uses DNA panels that look at a limited number of mutations. If the initial panel identifies only one mutation, in many cases the screen will be considered positive and the infant will be referred for diagnostic testing (a sweat test). However, in some cases, newborn screening labs may first perform further sequencing to determine whether a second mutation is present. Sequencing has also been proposed for second-tier screening for severe combined immunodeficiency (SCID).

NGS could potentially make an important contribution to second-tier newborn screening by enabling sequencing of the entire coding region of a gene (or genes) of interest, rather than being restricted to a limited number of pre-specified mutations. Thus, all mutations present in the coding region could be identified at once in second-tier screening, which might facilitate diagnostic testing in some cases.

A recent paper describes a method that takes this idea further, by using NGS to sequence 126 genes that together comprise the majority of the genes currently implicated in most newborn screening conditions (except hearing loss, critical congenital heart defects, and severe combined immunodeficiency). The method then uses a computer program to “filter” the sequence information so that only genetic variants potentially related to one of the specified diseases are reported. The researchers were able to use dried blood spots for some but not all of the targeted sequencing tests.

Although this method is suggested by the researchers as a second-tier screening test, it might someday be considered as a first-tier test for many of the newborn screening conditions. If all genes associated with the genetic conditions evaluated through newborn screening were sequenced at once, in a time- and cost-efficient manner, the need for follow-up testing might be greatly reduced. However, challenges remain before this strategy could be used in newborn screening, and further studies are needed. The recent study was not population-based, but instead used samples from children known to be affected by specific newborn screening disorders. Furthermore, while the technique worked well to identify potentially harmful sequence changes in the targeted 126 genes, this information alone only provided the correct diagnosis for 75% of the cases; for the rest, information about the child’s clinical condition was needed to complete the diagnosis.

Currently, NGS is expensive, although prices continue to drop as the technology improves. The need for rapid screening test results and the type of sample available for analysis—dried blood spots might not currently work in some cases—are also limitations that will potentially change as the technology evolves. More information will be needed on the sensitivity (the percentage of affected infants that are picked up by the screen) and specificity (the percentage of unaffected children that are not picked up by the screen) of NGS for the different newborn screening conditions. Tandem mass spectrometry measures the concentrations of metabolites in the blood, so that concentrations that are out of range (too high or too low) are detected directly without needing to know the mechanism or mutations involved. Additionally, metabolite concentrations that are out of range might reflect rare conditions not detectable by mutation analysis, but that require prompt treatment, such as vitamin B12 deficiency of the newborn. Thus, tandem mass spectrometry, at least for now, provides a more accurate way to detect inborn errors of metabolism. Most state newborn screening laboratories do not currently use NGS, so introducing this technology would require purchasing new equipment, hiring or training staff, and other considerations. Another challenge would be the large amount of data generated by NGS. Currently, newborn screening data systems deal with a much smaller set of data points for each screening test performed. For example, for cystic fibrosis screening, state laboratories would store the IRT concentration, the cutoff, and the results of the DNA mutation panel if second-tier screening was done. The informatics capabilities needed to store and analyze sequencing data would be much greater.

Some have suggested that NGS could be used to sequence the whole genome or exome of every infant to identify all the genetic conditions that may affect the infant now or in the future, rather than just focusing on the disorders included in the recommended newborn screening panel.  Four NIH-funded research projects are currently underway to find out whether whole-genome or exome sequencing of newborns can provide a useful resource of information to inform healthcare decisions throughout infancy and childhood (also see commentary in Nature). However, many experts and commentators urge caution, pointing out that even in current newborn screening programs there can be problems with interpreting some of the results of screening, and with our current state of knowledge, many of the results of genome or exome sequencing are not yet interpretable.  Current professional guidelines recommend that NGS should not be used as a first-tier newborn screening test.

For now, newborn screening using whole-genome or exome sequencing raises more questions than it answers. When state laboratories started using tandem mass spectrometry for newborn screening, some raised concern that the technology was driving the screening and cautioned that the ability to screen for a condition was not reason alone to report the results. Newborn screening using whole-genome or exome sequencing will face many of these same issues. The time to think about how to resolve them is now.