A new study shows that children born preterm who are later diagnosed with autism often present with more extensive support needs and a higher number of co-occurring conditions than autistic children born at full term. Surprisingly, however, the researchers found no differences in genetic variants across the genome, nor in specific genes already linked to autism, between the groups - a result that contradicted their initial hypothesis.
The study was conducted at KIND (Center of Neurodevelopmental Disorders at Karolinska Institutet) and published in October 2025 in the journal Genome Medicine.
"We did not observe any genetic differences between preterm and full-term autistic children, which was unexpected. We initially thought that preterm children might show fewer of the genetic factors associated with autism, as their early birth can be viewed as an environmental factor", says Yali Zhang , doctoral student at Tammimies research group at KIND and first author of the study.
The study did, however, reveal additive effects of genetics and prematurity. In addition, the researchers identified an increased number of new genetic variants - so-called de novo variants - in autistic preterm children compared with preterm children without autism.
AI-based tools may improve early support
The researchers also developed a machine learning model that, based on clinical and genetic data, could estimate the likelihood of autism in preterm infants already at or shortly after birth.
"AI tools hold great potential for predicting various conditions in healthcare - not to assign definitive labels, but to give children the best possible opportunities for early support", says Yali Zhang. "Our findings suggest that prematurity, in combination with genetic background, may justify closer follow-up and tailored care".
The research group now aims to validate their findings in larger, population-based cohorts with genetic and long-term health data, and hopes to further develop early screening tools for preterm children in clinical settings.
The full study
Prematurity and genetic liability for autism spectrum disorder.
Zhang Y, Yahia A, Sandin S, Åden U, Tammimies K
Genome Med 2025 Oct;17(1):108
