When a Newborn Has a Stroke: What Four New Studies Say About the Years Ahead

Research published in 2025 and 2026 on how children develop after a stroke around the time of birth — and why some difficulties only appear at school

Strokes happen to newborn babies far more often than most people realise, usually in healthy full-term infants whose first sign is a seizure in the first days of life. Four studies published between November 2025 and May 2026 followed these children into the school years. Almost all survive and almost none have a second stroke — but around four in ten have a lasting difficulty, and many of those difficulties involve language and vision rather than movement, and only become visible years later.

A problem that has always been hard for families to plan for

For most of the past century, a stroke in a newborn was something recognised only in hindsight — often years afterwards, when a toddler consistently favoured one hand and a scan finally revealed an old injury on the opposite side of the brain. The arrival of MRI changed that. Today, a term baby who has seizures in the first two or three days of life will usually have a scan within hours, and a blocked artery in the brain will be seen while the baby is still in hospital. It has also allowed researchers to count how often each type of perinatal stroke actually happens — work done across a whole Canadian province showed the condition is considerably commoner than its low public profile suggests [1].

What earlier detection has not yet delivered is a good answer to the question families ask first. The paradox of newborn stroke, described in a landmark review more than a decade ago, is that the baby almost always survives and almost never has another stroke, yet a substantial minority of children grow up with a lasting difficulty [2]. Studies from Switzerland established the typical picture — an apparently well term baby, seizures on day two or three, and a wide range of later outcomes [3]. Researchers have looked hard for causes and found only weak clues: a first pregnancy, pre-eclampsia, infection around the time of birth, an emergency caesarean. Each of these is common, and most babies whose mothers experienced them are entirely fine, so none of them can be used to predict or prevent anything [4]. Treatment in the first days is mostly about supporting the baby and controlling seizures; blood-thinning medication is used only in particular situations [5]. So families have often left hospital with a diagnosis, a good chance of survival, and very little else to hold on to.

What the four new studies looked at

The first study comes from Italy, where hospitals across the country have contributed to a national register of babies and children with blood-clotting problems since 2007 [6]. The researchers pulled out every baby whose stroke occurred in the first 28 days of life — 181 babies in all — and followed what happened to them, on average until they were about two years old.

The second study, from Australia and Switzerland, took a different approach [7]. Instead of asking simply where the damage was, it asked which brain networks the damage sat inside. Researchers used the scans 85 newborns already had, and compared them with a large map of normal newborn brain wiring built from 518 healthy babies scanned in London. The idea is that a brain works as a set of connected circuits, so an injury in one place can disturb regions far away — rather like a fault at one point in a rail network delaying trains at stations nowhere near it.

The third and fourth studies picked the story up much later. A French team assessed spoken language in 70 seven-year-olds who had had a newborn stroke [8]. A Canadian team tested visual perception and reading in 26 children aged 7 to 18, comparing them with 27 children of the same age and sex who had not had a stroke [9].

What they found

The Italian numbers are the ones most useful to a family at the bedside. Out of 181 babies, one died in hospital and none died afterwards. One baby had a second stroke while still in hospital, and none had another during follow-up. About eight in ten babies were first noticed because of a seizure. At discharge, 43 out of every 100 babies had something abnormal on their neurological examination, and by the last check-up around 39 in 100 still did — most often affecting muscle tone or movement. Twelve in every 100 children went on to develop epilepsy, and nearly nine out of ten of those children had already had seizures as newborns.

The study also identified which early signs went with a harder road ahead. The single most useful one was simply the examination the doctors performed before the baby went home: babies whose examination was abnormal at discharge had roughly seven times the odds of still having difficulties at follow-up. Babies who had needed a breathing machine, who had a longer hospital stay of 19 days or more, or whose scan showed changes in the brainstem were more likely to develop epilepsy later.

The imaging study found that a third of the 85 children developed cerebral palsy — a condition affecting movement and posture — diagnosed at around two years of age. In those children, the stroke was connected to a wide set of brain regions, including several that have nothing obvious to do with movement: the memory and emotion structures deep in the temporal lobe, parts of the frontal lobe, and the cerebellum at the back of the brain. The cerebellum is a telling detail, because it cannot itself be damaged by this type of stroke. Its appearance in the results is evidence that the effects of the injury spread through the brain's wiring rather than staying put.

The two school-age studies found what the earlier ones could not see. Among the seven-year-olds in France, 56% — more than half — had spoken-language skills below what would be expected for their age. Importantly, none of them had lost the ability to speak, and all had usable everyday language. The difficulties fell into four patterns: a large group with normal language, a small group with severe difficulties in speech sounds and sentence structure, a group whose main problem was understanding language, and a group with borderline scores across the board. The most striking finding concerned who got help. Every child whose difficulty was audible — trouble producing sounds and sentences — was receiving speech and language therapy. But only 4 of the 11 children whose difficulty was in understanding were getting any. Help was reaching the children whose problems could be heard, not the children whose problems were worst.

The Canadian study found something similar in a different domain. Children who had had a newborn stroke scored lower on a test of visual perception, took about 41 seconds longer to complete a visual search task, and read at a lower level than their peers. And here is the part that matters most: it made no difference whether the child had any movement difficulty at all. The children who walked and used both hands normally were struggling just as much as those who did not.

What this means for families

The first thing to take from all this is reassuring and should not be lost among the rest: a newborn stroke is very rarely fatal, and it very rarely happens twice. Families sometimes spend years quietly braced for a second event that, on this evidence, is not coming.

The second is that the years after discharge matter as much as the days in hospital. The pattern these studies describe — normal-looking preschooler, difficulties that surface at school — is not a sign that something new has gone wrong. It is what happens when a child who has had an early brain injury starts being asked to do harder things: follow a teacher's instructions, learn to read, keep up in a noisy classroom. Earlier long-term studies had already hinted at exactly this [10], [11]; these newer ones show it in detail.

Practically, that means it is worth asking for follow-up that continues into the school years, and worth asking specifically about language comprehension and about vision and reading — not only about walking and hand use. A general developmental check may pass a child whose real difficulty is understanding what is said to them. A commentator writing alongside the French study made this point directly: the children being overlooked are the ones whose difficulties are quiet [12]. If your child is having trouble at school and has been told their movement is fine, that is not a reason to accept that nothing is wrong.

It is also fair to know the limits of these studies. None of them tested a treatment; they observed what happened. Three of the four followed children at a single set of centres in one country, and the two school-age studies were small — 70 and 26 children. And none of them followed the same children from the newborn scan all the way to age seven, so the link between what the early MRI shows and what happens at school is still being pieced together.

What researchers are working on next

Three directions are taking shape. The first is turning the network-based scan analysis into something that gives a prediction for an individual baby rather than a pattern across a group — and making it predict language and visual difficulties, not only movement. Movement problems can already be identified early, in the first months, using MRI together with a careful assessment of how a baby moves spontaneously [13]; the outcomes with no early warning sign are precisely the ones these new studies describe. The second is better and earlier detection of quiet difficulties, so that referral for therapy depends on testing rather than on how noticeable a problem happens to be. The third is treatment. A team in the Netherlands has given stem cells through the nose to newborns in the days after a stroke, showing that the approach is feasible and appears safe [14]. Whether it helps is not yet known, and larger trials are needed. Taken together, these four studies make a strong argument that when those trials are designed, success should be measured by more than whether a child can use both hands.

References

  1. Dunbar M, Mineyko A, Hill M, et al. Population Based Birth Prevalence of Disease-Specific Perinatal Stroke. Pediatrics. 2020;146(5):e2020013201. doi:10.1542/peds.2020-013201 ↩
  2. Kirton A, deVeber G. Life After Perinatal Stroke. Stroke. 2013;44(11):3265–3271. doi:10.1161/STROKEAHA.113.000739 ↩
  3. Grunt S, Mazenauer L, Buerki SE, et al. Incidence and Outcomes of Symptomatic Neonatal Arterial Ischemic Stroke. Pediatrics. 2015;135(5):e1220–e1228. doi:10.1542/peds.2014-1520 ↩
  4. Lee J, Croen LA, Backstrand KH, et al. Maternal and infant characteristics associated with perinatal arterial stroke in the infant. JAMA. 2005;293(6):723–729. doi:10.1001/jama.293.6.723 ↩
  5. Ferriero DM, Fullerton HJ, Bernard TJ, et al. Management of Stroke in Neonates and Children: A Scientific Statement From the American Heart Association/American Stroke Association. Stroke. 2019;50(3):e51–e96. doi:10.1161/STR.0000000000000183 ↩
  6. Sartori S, Nosadini M, Brigiari G, et al. Outcome and epilepsy following neonatal stroke in the Italian Registry of Infantile Thrombosis. Eur J Pediatr. 2026;185(6):352. doi:10.1007/s00431-026-06912-8 ↩
  7. Kelly CE, Chen J, Beare R, et al. Predicting Outcome After Newborn Stroke: A Lesion Network Mapping Study Leveraging Large-Scale Data. Stroke. 2026;57(6):1706–1716. doi:10.1161/STROKEAHA.125.052941 ↩
  8. Drutel L, Dardier V, Avoyan L, Hertz-Pannier L, Dinomais M. Oral language profiles and associated factors in children after neonatal arterial ischaemic stroke. Dev Med Child Neurol. 2026;68(8):1105–1116. doi:10.1111/dmcn.70132 ↩
  9. Maiani M, Lahey N, Bardhi M, et al. Visuo-perceptual Abilities in Children With Perinatal Arterial Ischemic Stroke and Associated White Matter Microstructure. J Child Neurol. 2026;41(6):840–850. doi:10.1177/08830738251394438 ↩
  10. Chabrier S, Peyric E, Drutel L, et al. Multimodal Outcome at 7 Years of Age after Neonatal Arterial Ischemic Stroke. J Pediatr. 2016;172:156–161.e3. doi:10.1016/j.jpeds.2016.01.069 ↩
  11. Wagenaar N, Martinez-Biarge M, van der Aa NE, et al. Neurodevelopment After Perinatal Arterial Ischemic Stroke. Pediatrics. 2018;142(3):e20174164. doi:10.1542/peds.2017-4164 ↩
  12. Westmacott R. Rethinking language outcomes after neonatal stroke. Dev Med Child Neurol. 2026;68(8):1026–1027. doi:10.1111/dmcn.70130 ↩
  13. Novak I, Morgan C, Adde L, et al. Early, Accurate Diagnosis and Early Intervention in Cerebral Palsy: Advances in Diagnosis and Treatment. JAMA Pediatr. 2017;171(9):897–907. doi:10.1001/jamapediatrics.2017.1689 ↩
  14. Baak LM, Wagenaar N, van der Aa NE, et al. Feasibility and safety of intranasally administered mesenchymal stromal cells after perinatal arterial ischaemic stroke in the Netherlands (PASSIoN): a first-in-human, open-label intervention study. Lancet Neurol. 2022;21(6):528–536. doi:10.1016/S1474-4422(22)00117-X00117-X) ↩