A recent study published in Translational Psychiatry suggests that the chemical makeup of blood in a newborn’s umbilical cord might offer early clues about their likelihood of developing conditions like autism or ADHD. By analyzing specific molecules related to cholesterol and amino acids, scientists discovered a pattern that provides evidence of increased risk for these developmental differences years before symptoms typically appear. This finding points toward a future where a simple blood test at birth could help identify children who might benefit from earlier support and intervention.
Childhood-onset neurodevelopmental disorders affect up to one in five children worldwide. These conditions encompass a variety of differences in brain development and function. The most common examples include attention-deficit hyperactivity disorder (ADHD), autism spectrum disorder (ASD), and specific delays in speech, learning, or motor coordination.
Diagnosing these conditions usually relies on observing a child’s behavior and developmental milestones over time. The average age for an initial diagnosis is around five and a half years old, though many children are diagnosed much later.
Polina Girchenko, a researcher at the University of Oulu and the University of Helsinki in Finland, noted that this timeline misses a highly sensitive developmental window. “Most neurodevelopmental disorders are only diagnosed after symptoms appear, often at school age, by which time the most important phase of brain development has already passed,” Girchenko said. “We wanted to know whether the risk of these disorders could be detected already at birth.”
The time spent in the womb is a deeply influential period for fetal brain development. The environment inside the uterus can shape how the nervous system forms and wires itself. The scientific framework for this concept suggests that health and disease outcomes in childhood and adulthood often originate from prenatal exposures.
To better understand this prenatal environment, scientists can look at the umbilical cord blood at the moment of birth. Cord blood contains a mix of substances transported from the mother through the placenta, as well as substances produced by the fetus itself. Among these substances are metabolites.
Metabolites are the small molecules left behind when the body builds tissue, breaks down food, or processes chemicals. Studying the complete set of these tiny molecules is known as metabolomics. Because cord blood acts as a chemical snapshot of the newborn’s physiological state, metabolomics can reveal detailed information about the biological processes happening just before birth.
Past research examining cord blood for signs of neurodevelopmental conditions has produced mixed and sometimes conflicting results. Many previous efforts only looked at a small handful of specific metabolites. Other projects focused strictly on autism or ADHD, missing a broader range of childhood developmental differences that frequently overlap.
To address these gaps, the authors of the current paper designed a broader approach. They wanted to see if a wide array of metabolic markers could predict any type of neurodevelopmental disorder. They also wanted to know if analyzing these molecules provided evidence beyond what is already known from standard early-life risk factors, such as maternal health or birth weight.
To explore this connection, the researchers analyzed data from an existing project called the PREDO study. This project initially tracked a cohort of pregnant women and their children born in Finland between 2006 and 2010. The researchers followed the 858 children included in this specific analysis from birth until they were teenagers, reaching a median age of about 14 years old.
At the time of each child’s birth, medical staff collected blood from the umbilical vein. The scientists stored this plasma and later tested it using a high-throughput technology that identifies and measures specific molecules using magnetic resonance. They looked at a targeted panel of 110 different metabolic measures.
These 110 measures covered major biological functions like fluid balance, inflammation, and the processing of fats and sugars. Over the following years, the researchers tracked the children using a highly reliable nationwide healthcare register to see who received a formal neurodevelopmental diagnosis. Out of the 858 children, 120 met the criteria for at least one neurodevelopmental disorder.
The authors then used statistical modeling to compare the cord blood of the children who later received diagnoses with the blood of the 738 children who did not. They identified a specific combination of 12 metabolic measures that were collectively associated with a higher likelihood of developing a neurodevelopmental disorder.
Seven of these 12 molecules were related to high-density lipoprotein, commonly known as HDL. In adult health, HDL is often referred to as “good cholesterol” for its role in heart health. In fetuses, however, HDL cholesterol is a necessary building block for cell membranes and for the production of hormones that support optimal brain function.
The remaining five markers included the total concentration of all fat-carrying particles in the blood, two types of fatty acids, and two amino acids. Amino acids are the basic chemical units that form proteins. The two specific amino acids identified in this pattern were alanine and histidine. The fatty acids included omega-6 fatty acids and a measure of fat unsaturation, which are involved in brain growth and signaling.
The scientists found that the levels of all 12 of these specific metabolites were consistently lower in the cord blood of the children who later developed a neurodevelopmental disorder. This pattern of lower metabolic markers was present across various specific diagnostic categories. It appeared in children with ADHD, speech disorders, learning difficulties, and autism.
The researchers even found that lower levels of these molecules correlated with milder, subclinical symptoms of developmental delay. These milder symptoms were reported by mothers on screening questionnaires when the children were toddlers.
Next, the authors checked if this metabolic pattern provided new predictive information. They compared the predictive power of the metabolites against established early-life risk factors. These standard risk factors included the mother’s age, education level, body mass index before pregnancy, and mental health during pregnancy. They also included the baby’s birth weight and gestational age.
Interestingly, many of the maternal risk factors, such as higher body mass index and prenatal depression, correlated with the exact same pattern of lower cord blood metabolites. This suggests that the mother’s health and environment might influence the availability of essential fats and amino acids for the developing fetus.
The standard risk factors alone could explain about 12.2 percent of the variance in whether a child developed a neurodevelopmental condition. When the researchers added the 12 cord blood metabolites to their statistical model, the predictive accuracy improved. The combined model accounted for 16.6 percent of the variance.
The researchers also looked at a smaller group of 737 mothers who had genetic data available. They calculated genetic risk scores to estimate the mother’s inherited likelihood of having ADHD and autism. Even when factoring in this maternal genetic risk, the cord blood metabolites still offered a modest improvement in predicting a child’s future diagnosis.
“Cord blood carries chemical signals that can indicate a child’s later risk of neurodevelopmental disorders, and these signals add information beyond what standard clinical and newborn screening data provide,” Girchenko told PsyPost. “This means one day, a blood test in connection with clinical information could help identify children who would benefit from early prevention.”
She added that this kind of proactive approach could eventually change the trajectory of a child’s development. “Even modest improvements in early detection can have significant public health effects, because the first years of life are when the brain is most responsive to intervention,” Girchenko explained. “Identifying risk at birth is the first step; the larger goal is to translate that knowledge into preventive actions that reduce the burden on children, families, schools, and healthcare systems.”
The findings from this project introduce an intriguing biological link, but there are several limitations to consider before these markers can be used in clinics. The authors emphasize that this research represents an early stage of scientific discovery.
“This is a proof-of-concept study,” Girchenko noted. “We have shown that the risk signal exists in cord blood and can be measured, but the biomarkers need to be independently verified before they can be used in any clinical screening program. This is not a diagnostic test.”
The original group of mothers in the PREDO study was largely recruited because they had a high clinical risk for pregnancy complications like pre-eclampsia. This means the sample had a higher rate of prenatal difficulties than the general population, which might skew the typical metabolic profile.
The statistical analyses also relied on the same group of children to both identify the metabolic markers and test their predictive power. In scientific research, testing a newly discovered pattern on an entirely separate group of people helps confirm its accuracy. Future projects will need to replicate these findings in different populations to verify that these 12 molecules are reliable predictors.
The study took place in a high-resource Nordic country. This limits how well the findings might apply to people in different geographic or economic environments. The healthcare systems, dietary habits, and environmental exposures in Finland might not reflect the daily realities of populations in other parts of the world.
It is important to note that these predictive models do not establish a direct cause and effect. Lower levels of these metabolites do not guarantee that a child will develop a neurodevelopmental disorder. Instead, they might simply act as a biological footprint of other processes happening during pregnancy, such as maternal stress, immune system activity, or nutritional deficits.
The researchers only analyzed a specific panel of 110 metabolic measures. Although this panel covered many major biological pathways, human blood contains thousands of distinct molecules. A more comprehensive metabolomic analysis could potentially uncover additional patterns or even stronger predictors that this limited panel missed.
Looking ahead, the research team aims to test their models in new settings. Girchenko said their next steps involve “validation in independent cohorts, ideally across different populations, and combining these metabolic biomarkers with other molecular and clinical data.”
Scientists might also explore whether maternal nutrition or lifestyle modifications during pregnancy could safely alter these metabolic markers. The authors note that previous clinical trials have tested whether supplementing children’s diets with certain fatty acids can reduce ADHD symptoms, suggesting that some metabolic pathways might be responsive to targeted interventions.
“The long-term aim is a verified newborn screen for neurodevelopmental risk, paired with targeted primary prevention strategies that can be offered before symptoms ever appear,” Girchenko said.
The study, “Metabolomic alterations in cord blood improve the prediction of childhood-onset neurodevelopmental disorders,” was authored by Polina Girchenko, Marius Lahti-Pulkkinen, Chenyao Ni, Jari Lahti, Li Tian, Aino Airikka, Eero Kajantie, and Katri Räikkönen.