From a Drop of Blood to a Genome: The Future of Newborn Screening

Barely aware of their surroundings, it begins with a prick of a needle and a drop of blood. In the past, the blood was placed on a petri dish. If Bacillus subtilis grew, it meant that the baby had Phenylketonuria (PKU) caused by a deficiency in phenylalanine hydroxylase [1]. When unrecognized, PKU results in high levels of phenylalanine that can cause brain damage; with the proper diet, these children can live a normal life. Since the pioneering work of Robert Guthrie, newborn screening has become an invaluable standard of care. 

Over the years, screening has expanded to include a wide range of genetic conditions, including inborn errors of metabolism, like PKU, cystic fibrosis, hemoglobinopathies, and endocrine disorders. With the development of genome sequencing, it is now possible to screen newborns for far more conditions, especially rare ones, that might not be clinically expressed until so advanced that treatment is less effective. 

Medicine advances too slowly for some, especially parental and public health advocates for children with rare genetic disorders. JAMA provides us with the initial report of the Genomic Uniform-screening Against Rare Disease in All Newborns (GUARDIAN) study, which seeks to demonstrate that genomic sequencing can replace our current methods. Let’s begin with the findings and then consider the hurdles from the lab to the real world. 

GUARDIAN

The study, conducted within the New York State Department of Health Newborn Screening program, extracted DNA from those dried blood spots used routinely since 1961. The testing was a supplement, not a substitute for current screening. Because not all genetic conditions can be treated, the study looked at 237 genes associated with 255 discrete conditions. 60% had established interventions, another 38% were associated with preventable seizures, and the remainder considered “treatable.”

“In my practice, I’ve seen many patients who’ve spent years going from doctor to doctor with symptoms that no one can explain. But by the time they receive a diagnosis, the window to best manage the disease has usually passed. … Families and pediatricians don’t need to go through those diagnostic odysseys anymore with the genomic technology we now have. We can make the diagnosis at birth.” 

- Wendy Chung MD, Chief of Pediatrics at Boston Children’s Hospital 

Negative results were shared with parents via phone and encrypted email. Positive results were delivered by phone, followed by an in-person visit with a medical geneticist, genetic counselor, and other specialists. Further clinical testing, including parental genetic testing, was conducted as needed during the visit. For purposes of the study:

• True positives: Variants confirmed by clinical testing or functional data.
• Presumptive positives: Screen-positive infants where diagnosis couldn’t be confirmed due to age.
• False positives: Cases where further testing ruled out the disease.

Between September 2022 and July 2023, there were 8617 newborns, and 5555 parents, 65%, were approached to participate. Four thousand consented to participation and are the cohort being reported. The most common reason for declining participation was a lack of interest, the perception that standard testing was sufficient, or feeling overwhelmed. Single percentages expressed concerns about genetic testing or privacy. The mean time for turnaround for a report was 49 days, but overtime turnaround was reduced to about one month. 

The study was not sufficiently populated to draw statistically significant information regarding true and false positives and negatives. With that caveat,

• 26 infants were found to be positive by conventional newborn screening. 8 of these newborns were positive by genomic testing.
• 147 (3.7%) of the 3982 newborns successfully sequenced screened positive, two-thirds for G6PD deficiency, an inborn error in metabolism that causes red blood cells to be fragile resulting in their earlier destruction. It is the most common enzyme deficiency anemia. [2] 79% were true positives, 4% were presumptive positives, and about 16% were false positives. 
• There were 110 true positive tests for conditions not screened by the current testing. These conditions, all treatable, included Severe Combined Immunodeficiency (SCID); long QT syndrome, an abnormality of conduction of heart signals; achondroplasia and hypochondroplasia, abnormalities of bone growth resulting in dwarfism;  and Wilson disease, causing abnormal accumulations of copper. 

Bringing the lab to the real world

These results are quite promising, and when the study gathers the remaining 6,000 participants and is sufficiently statistically powered, we will have a better idea of how it measures up as a replacement screening test for newborns. But even if the testing is better, logistical hurdles remain.

The results from these tests were overread by humans. While 94% did not require further expert evaluation, 6% required a deeper analysis. Newborns of African ancestry were more frequently positive than those of European ancestry and required more “manual review.” The manual review adds time and money costs, reducing testing scalability. In addition, greater positive screens add to the necessary downstream costs. However, a back-of-the-envelope analysis would suggest that the cost of additional counseling will more than be paid back by more newborns becoming healthier children and adults.

“These costs need to be weighed against the costs incurred if a child gets sick with a condition that could have been treated if caught earlier, and the value of saving lives…. Ultimately, it’s a matter of who will pay for it. When you realize that genomic screening can detect so many more conditions, prevent more illness and save lives, the extra cost may be worth it.” 

- Joshua Milner, professor of Pediatrics, Director of Allergy/Immunology and Rheumatology at Columbia University 

There is also the significant time required to carry out the DNA extractions, genomic testing, analysis, and reporting. Currently, at best, we are looking at a month’s wait – much longer than anxious parents might tolerate. And while on the topic of anxiety, will false positives have significant negative impacts on parental mental health? The researchers suggest that traditional and genomic newborn testing be applied synergistically, but now genomic testing is no longer a substitute but an extension of newborn screening. 

Lastly, some issues might be more deemed technical. What genomic diseases should be included? Do we include the untreatable, and if so, what are the repercussions? Individuals at risk for Huntington’s Disease often avoid diagnostic testing until they are symptomatic – not knowing can give comfort in impossible situations. Who will be the arbiter of what to include, and what criteria will be employed to separate what is often a spectrum of risk into a yes or no categorization?

The GUARDIAN study's promising early results point to a future where genomic screening could catch conditions that today's methods miss. Yet, real-world hurdles—manual review, false positives, long wait times, and ethical dilemmas—remain significant challenges to widespread adoption. As science advances, so must our approach. Whether genomic screening will be the solution or just another tool in the toolbox is a question that only time (and more data) will answer.

[1] The petri dishes substrate, agar, contains a phenylalanine antagonist. The excess phenylalanine in the baby with PKU overcomes the inhibition, resulting in bacterial growth. Screening began in earnest in 1951, and within 2 years, 39 cases of PKU were identified in 400,000 newborns tested. 

[2] In New York State, G6PD testing is not part of newborn screening, being performed selectively “for infants with hemolytic anemia, hemolytic jaundice, early-onset increasing neonatal jaundice persisting beyond the first week of life admitted to hospital for jaundice following discharge, or familial or population risk for G6PD deficiency.” 

Source: Expanded Newborn Screening Using Genome Sequencing for Early Actionable Conditions JAMA DOI: 10.1001/jama.2024.19662