A Simple Blood Test That May One Day Help Predict a Baby's Recovery After a Difficult Birth

New research building on the HEAL trial suggests a protein in a newborn's blood could help forecast long-term outcomes after brain injury at birth

When a baby is born after a difficult delivery and doctors diagnose a brain injury called hypoxic-ischemic encephalopathy, parents are almost always asked to wait for answers about their child's future. New research suggests that a simple blood test, taken in the first two days of life, might one day help doctors give families a clearer picture much sooner — potentially as early as day two, instead of a week or more later.

Why Predicting a Baby's Future Matters So Much, So Early

Hypoxic-ischemic encephalopathy, usually shortened to HIE, happens when a baby's brain does not get enough oxygen and blood flow around the time of birth — sometimes because of a problem with the umbilical cord, the placenta, or a difficult labor. It is one of the more common serious brain injuries seen in full-term newborns, and it can lead to cerebral palsy (a condition that affects movement and muscle coordination) or to delays in thinking, learning, and language as the child grows. For the parents living through it, the days after diagnosis are often the hardest: their baby is connected to monitors in the neonatal intensive care unit (NICU), receiving a cooling treatment meant to protect the brain, and no one can yet tell them how their child's story will unfold. Being able to answer that question earlier and more confidently — even partially — would change how families plan, how doctors counsel them, and how quickly a baby can be connected with therapies and support services that make the biggest difference when started early. It could also simply ease some of the emotional weight of those first NICU days, replacing an open-ended unknown with real, if imperfect, information a little sooner.

How Families Used to Face This "Wait and See" Period

A team of researchers recently set out to test whether a simple blood sample, drawn in a baby's first days of life, could answer that question sooner — using data from a large research effort called the HEAL trial [1]. To understand why that question matters so much, it helps to know how far, and how incompletely, treatment for HIE has already come. For a long time, there was no treatment at all for HIE beyond supportive care — keeping a baby warm, stable, and safe while the brain injury ran its course. That changed in the mid-2000s, when three major research studies — a large U.S. trial [2], a UK-led international trial called CoolCap [3], and a second UK-based trial called TOBY [4] — showed that cooling a newborn's body to about 33.5°C for three days, starting soon after birth, could protect the brain and improve outcomes. This "therapeutic hypothermia," or cooling therapy, quickly became the standard treatment offered in NICUs around the world, and it remains standard today. But cooling is not a cure: even with it, a meaningful number of babies with moderate or severe HIE still go on to develop cerebral palsy or learning delays, and HIE remains a real and ongoing cause of long-term disability even in modern, well-resourced hospitals [5]. Because of that gap, researchers went looking for something to add to cooling that might protect the brain even further. One promising candidate was erythropoietin, a naturally occurring hormone best known for helping the body make red blood cells, which laboratory studies suggested might also calm inflammation and protect brain cells after an injury like this.

That idea was tested in a large, carefully designed study called the High-dose Erythropoietin for Asphyxia and EncephaLopathy trial, known as HEAL [6]. Conducted at 17 hospital sites across the United States between 2017 and 2019, HEAL gave several hundred newborns with HIE either high-dose erythropoietin or a placebo, on top of the standard cooling treatment. When the results came in, they were disappointing: babies who received erythropoietin did not do any better than those who received placebo — about half of each group experienced death or a significant developmental problem by age two — and the erythropoietin group actually had somewhat more serious side effects. So while HEAL did not find a new treatment, it left behind something valuable: blood samples that had been carefully collected from these babies in the first days of life, along with detailed brain scans and, years later, careful assessments of how each child was doing. That leftover treasure trove is exactly what the new study made use of.

What the New Study Found

A research team led by Dr. Danielle Clifford went back to those saved blood samples from 153 of the babies in the HEAL study — the ones who had samples from all three collection points (soon after birth, day 2, and day 4), a brain MRI, and a follow-up developmental assessment at age two using a standard testing tool called the Bayley Scales [7]. They tested each sample for 27 different proteins that earlier research had already linked to brain injury and inflammation [8] — including a separate group of researchers who had already found that babies with higher levels of tau during cooling treatment were more likely to have severe brain injury and trouble regulating blood flow to the brain [9] — to see whether any of them, measured so early, could predict which babies would go on to develop cerebral palsy or delays in thinking and language. As a comparison, they also used a detailed scoring system that doctors already use to grade brain-injury severity on MRI scans [10], done on day 4 to 6 of life.

The standout result involved a protein called tau, which is released when brain cells are injured. When measured on day 2 of life, tau predicted which babies would later be diagnosed with cerebral palsy remarkably well — correctly flagging about 9 out of 10 affected babies (88%) and correctly clearing about 9 out of 10 unaffected babies (91%). Put another way, this simple blood test performed about as well as the detailed brain MRI that doctors currently rely on — except the MRI cannot usually be done until day 4 to 6 of life, while the blood test could be run two to four days sooner. When researchers combined the blood test and the MRI together, the prediction became even more accurate still. A handful of other proteins in the blood — with names like GFAP, IL-13, and VEGF — also showed some connection to cerebral palsy risk, though none performed quite as well as tau.

Predicting delays in thinking, learning, and language proved much harder. For milder delays, none of the blood tests performed particularly well on their own. For more significant delays, tau again did best among the individual proteins tested, correctly predicting the outcome in roughly two-thirds to just over two-thirds of cases on its own, rising to about three-quarters of cases when combined with the MRI — helpful information, but a far less precise answer than the near-perfect performance seen for cerebral palsy. The plain-language takeaway is that this kind of blood test may end up being much more useful for predicting movement-related outcomes like cerebral palsy than for predicting learning and language outcomes, which appear to be harder to forecast from blood alone with the tools tested so far.

What This Means for Families Today — and What Comes Next

It is important to be clear about what this research does and does not mean for a family in the NICU right now. This blood test is not yet available for clinical use anywhere. The specific protein levels used in this study came from one research study, one laboratory setup, and one group of babies, and researchers will need to repeat and confirm these findings in other hospitals and other groups of children before any doctor could responsibly use tau levels to counsel a family about their baby's future. It is also worth knowing that most of the babies in this study were cared for in major U.S. academic hospitals and were more often white and from families with higher levels of education, so it is not yet known whether the same blood test would work as well for babies from more diverse backgrounds or in different countries and healthcare settings. Because the number of babies who developed cerebral palsy in this study was small — only 18 out of 153 — the impressively high numbers reported for that prediction, while genuinely promising, will need to be confirmed in a larger group before anyone can be fully confident in them.

What this study does offer is genuine hope for the future of prognosis after a difficult birth. The researchers point out that some of the same proteins studied here, including GFAP, are already being developed as quick, portable tests — the kind that could eventually be used at the bedside rather than sent to a specialized lab — which could one day make this kind of early information available even in hospitals without easy access to brain MRI, including in parts of the world where that technology is scarce. Before that becomes a reality, researchers are working on confirming these results in larger and more diverse groups of babies, following children for longer than two years to see how they do once they reach school age, and testing whether combinations of several blood proteins together can predict thinking and language outcomes as well as tau already predicts cerebral palsy. Researchers are also interested in whether identifying high-risk infants earlier could eventually help match babies with newer, experimental therapies — such as stem-cell treatments now being studied — that might be given after the standard cooling window has already closed. For families facing an HIE diagnosis today, the message from this research is a hopeful one: scientists are actively working to shrink the "wait and see" period that has defined this experience for decades, replacing uncertainty, a little earlier each year, with real information.

References

  1. Clifford D, Vaughan CL, Costelloe SJ, et al. Blood-Based Biomarkers Predict Cerebral Palsy and Cognitive Delay in Hypoxic-Ischemic Encephalopathy: A Secondary Analysis of the High-Dose Erythropoietin for Asphyxia and Encephalopathy Randomized Controlled Trial. J Pediatr. 2026;297:115216. doi:10.1016/j.jpeds.2026.115216 ↩
  2. Shankaran S, Laptook AR, Ehrenkranz RA, et al; National Institute of Child Health and Human Development Neonatal Research Network. Whole-Body Hypothermia for Neonates with Hypoxic-Ischemic Encephalopathy. N Engl J Med. 2005;353(15):1574-1584. doi:10.1056/NEJMcps050929 ↩
  3. Gluckman PD, Wyatt JS, Azzopardi D, et al. Selective head cooling with mild systemic hypothermia after neonatal encephalopathy: multicentre randomised trial. Lancet. 2005;365(9460):663-670. doi:10.1016/S0140-6736(05)17946-X17946-X) ↩
  4. Azzopardi DV, Strohm B, Edwards AD, et al; TOBY Study Group. Moderate Hypothermia to Treat Perinatal Asphyxial Encephalopathy. N Engl J Med. 2009;361(14):1349-1358. doi:10.1056/NEJMoa0900854 ↩
  5. Cornet MC, Kuzniewicz M, Scheffler A, et al. Perinatal Hypoxic-Ischemic Encephalopathy: Incidence over Time Within a Modern US Birth Cohort. Pediatr Neurol. 2023;149:145-150. doi:10.1016/j.pediatrneurol.2023.08.037 ↩
  6. Wu YW, Comstock BA, Gonzalez FF, et al; for the HEAL Consortium. Trial of Erythropoietin for Hypoxic-Ischemic Encephalopathy in Newborns. N Engl J Med. 2022;387(2):148-159. doi:10.1056/NEJMoa2119660 ↩
  7. Bayley N. Bayley Scales of Infant and Toddler Development, 3rd ed. San Antonio, TX: Harcourt Assessment; 2006. ↩
  8. Yang Z, Xu H, Sura L, et al. Combined GFAP, NFL, Tau, and UCH-L1 panel increases prediction of outcomes in neonatal encephalopathy. Pediatr Res. 2023;93(5):1199-1207. doi:10.1038/s41390-022-01994-0 ↩
  9. Li R, Lee JK, Govindan RB, et al. Plasma Biomarkers of Evolving Encephalopathy and Brain Injury in Neonates with Hypoxic-Ischemic Encephalopathy. J Pediatr. 2023;252:146-153.e2. doi:10.1016/j.jpeds.2022.07.046 ↩
  10. Shankaran S, Barnes PD, Hintz SR, et al; Eunice Kennedy Shriver NICHD Neonatal Research Network. Brain injury following trial of hypothermia for neonatal hypoxic-ischaemic encephalopathy. Arch Dis Child Fetal Neonatal Ed. 2012;97(6):F398-F404. doi:10.1136/archdischild-2011-301524 ↩