Premature birth does not directly cause autism, but it significantly increases the risk. Babies born before 37 weeks are diagnosed with autism at roughly 2.1% compared to 1.4% for those born at full term, and the earlier a baby arrives, the higher that risk climbs. For the most premature infants, born between 22 and 27 weeks, the rate reaches 6.1%, more than four times the full-term rate.
The relationship is real and consistent across large studies, but it’s not a simple cause-and-effect story. Prematurity appears to be one of several factors that can tip the scales, particularly when genetic vulnerability is already present.
How Risk Rises With Earlier Birth
A Swedish national cohort study tracking over four million births found a clear, graded pattern: the earlier a baby is born, the more likely an autism diagnosis later in childhood. The numbers break down by gestational age like this:
- Extremely preterm (22 to 27 weeks): 6.1% diagnosed with autism
- Very to moderate preterm (28 to 33 weeks): 2.6%
- Late preterm (34 to 36 weeks): 1.9%
- Early term (37 to 38 weeks): 1.6%
- Full term (39 to 41 weeks): 1.4%
Even early term birth, just one to two weeks before the standard “full term” window, carried a slightly elevated risk. This gradient suggests that brain maturation in the final weeks of pregnancy plays a meaningful role. The association held for both boys and girls, though boys born preterm face a notably higher risk overall. A meta-analysis of population-based studies found that male sex among preterm infants was associated with roughly three times the risk of autism compared to females.
Low Birth Weight Adds Its Own Risk
Birth weight matters independently of how early a baby arrives. Babies born weighing less than 1,500 grams (about 3.3 pounds) had roughly three times the odds of an autism diagnosis compared to normal-weight newborns, even after researchers accounted for gestational age, maternal age, smoking, and other factors. Moderately low birth weight (under 2,500 grams) still carried about 60% higher odds.
Being small for gestational age, meaning the baby weighed less than expected for how far along the pregnancy was, also raised the risk. Among the most premature babies (23 to 31 weeks), those in the smallest 5% for their gestational age had 60% higher odds of autism than appropriately grown preterm babies at the same gestational age. Interestingly, in that same very premature group, babies who were larger than expected actually showed a reduced risk, roughly half the odds. This pattern didn’t hold for babies born closer to term, suggesting that growth restriction in the earliest weeks of development is particularly impactful.
What Happens in the Premature Brain
The last trimester of pregnancy is a critical window for brain wiring. Neurons are forming connections, insulation around nerve fibers is developing, and the brain’s sensory and social processing networks are being organized. When a baby is born early, this process continues outside the womb in a radically different environment.
Brain imaging studies of newborns show that preterm infants have atypical patterns of dynamic connectivity, the way different brain regions coordinate their activity in real time. Specifically, preterm babies show lower overall synchronization across brain networks at birth, and this reduced synchronization is linked to higher scores on early autism screening measures at 18 months. The disruption is especially notable in sensory processing networks and in the connections between frontal and parietal regions of the brain, areas that later support social attention and flexible thinking.
These aren’t just statistical curiosities. Newborns who went on to show more atypical social and repetitive behaviors as toddlers had measurably different brain connectivity patterns within days of birth, suggesting the neurological groundwork is laid very early.
The Role of Inflammation and Pain
Premature infants in neonatal intensive care units undergo an average of ten procedures per day, many without pain relief. This matters more than it might seem. In the neonatal period, the brain is highly plastic, and repeated painful or stressful experiences trigger inflammatory responses that can alter brain development.
Inflammatory pain in newborns causes a cascade of effects: increased levels of inflammatory signaling molecules in both the blood and brain, activation of the brain’s immune cells, and cell death in the cortex and hippocampus. In animal models, these early inflammatory insults led to impaired nerve fiber growth, loss of protective insulation around nerve fibers, and reduced production of new brain cells, all within two weeks of the initial injury. The animals later displayed behaviors associated with autism.
When inflammation was blocked with anti-inflammatory treatment at the time of the painful stimulus, both the brain changes and the behavioral differences were prevented. This suggests that inflammation itself, not just prematurity, is a key part of the pathway. Premature infants are also more likely to experience infections, breathing difficulties, and surgical procedures, all of which generate inflammatory responses during a period when the brain is exceptionally vulnerable.
Genetics, Prematurity, or Both
One of the trickiest questions is whether prematurity independently raises autism risk or whether some babies are genetically predisposed to both being born early and developing autism. The answer appears to be: it’s a combination.
A 2025 study in Genome Medicine found that preterm and full-term children with autism carried similar overall levels of common genetic variants associated with autism. In other words, preterm birth didn’t seem to select for children with a higher genetic “load” for autism through common inherited variants. However, preterm children with autism did show a significantly higher rate of rare, spontaneous genetic mutations (called de novo variants) compared to preterm children without autism. These rare mutations may make it more likely that a preterm child crosses the diagnostic threshold.
The most striking finding was how these factors combine. Higher genetic risk scores for autism, preterm birth, and male sex were additive. A boy born preterm with a high genetic risk score had a predicted probability of autism diagnosis approaching 90%. Each factor on its own raised risk modestly, but together they compounded dramatically. Preterm children with autism also showed more severe outcomes overall, including more co-occurring conditions, even though their underlying genetic profiles for autism were similar to those of full-term children with autism. This suggests prematurity adds developmental challenges on top of whatever genetic vulnerability exists.
How Autism Looks in Preterm Children
Autism in children born preterm can be harder to identify because prematurity itself causes developmental differences that overlap with autism traits. Preterm children commonly show delays in executive function (planning, attention, impulse control) and social development, even without autism. This overlap can delay diagnosis or lead to initial misattribution of symptoms.
Some features help distinguish the two. Stereotypic and self-injurious behaviors, along with rigid, repetitive routines, are characteristic of autism but not typical of prematurity-related developmental delays on their own. Eye-tracking studies have shown that preterm toddlers with autism fixate less on faces compared to both typically developing children and preterm children with other developmental delays, mirroring the pattern seen in full-term children with autism. Preterm children without autism, even those with other delays, look at faces in ways that resemble full-term typically developing children.
Motor delays in the first years of life can also be a clue. In children with autism, early motor difficulties tend to predict later social communication problems, a connection that hasn’t been found in preterm children without autism.
Screening Preterm Infants
Because of the elevated risk, developmental follow-up for preterm infants typically includes autism-specific screening. The most widely used tool is the M-CHAT, a 23-item yes/no questionnaire for parents, optimally given around 24 months of age. For preterm populations, research supports combining the M-CHAT with broader developmental screens that capture communication, sensory processing, and social engagement as early as 6 months.
Using multiple screening tools matters in this population. A single questionnaire can miss cases or produce false positives because prematurity-related delays can mimic autism traits. Studies of very low birth weight infants have found that combining a targeted autism screen with a broader communication checklist and a sensory processing measure gives a more reliable picture. Children who screen positive on any tool are then referred for comprehensive diagnostic evaluation, which can distinguish autism from other developmental effects of prematurity.
Screening at around 18 to 24 months (adjusted for prematurity) catches many cases, but some children born very early may not show clear autism-specific signs until later, particularly if their overall development is delayed. Ongoing monitoring through preschool age is standard for children born before 28 weeks or weighing under 1,500 grams at birth.