Babies born a week early have a “greater risk of autism,” The Daily Telegraph has claimed.
The news is based on research that found that babies delivered between 37 and 39 weeks had a higher chance of being later diagnosed with special educational needs (SEN) such as autism and dyslexia than babies delivered at the full term of 40 weeks. However, pregnant women should not be alarmed by this research or media reports as the risk of developing SEN was still relatively low in early babies, with delivery a week early linked to only around three extra cases per 1,000 births. Also, the study looked at special educational needs as a whole, meaning it did not report any specific rise in the risk of autism.
The findings have important implications for the timing of elective caesarean deliveries, with the researchers suggesting that, ideally, such deliveries should be postponed until 40 weeks when possible.
Where did the story come from?
The study was carried out by researchers from the Section of Public Health at the University of Glasgow, the Department of Obstetrics and Gynaecology at Cambridge University, the Information Services Division of NHS Scotland in Edinburgh and the Rosie Hospital, Cambridge. It was funded by NHS Health Scotland and published in the peer-reviewed journal Public Library of Science Medicine.
Both the Telegraph and the Daily Mail reported on this study. Their headlines and lead paragraphs, which emphasised the risk for babies born one week early, seem unnecessarily alarmist given the low increase in individual risk for babies born at 37-39 weeks. The Mail did report the lead author’s advice that women having planned caesareans should not panic, and that the chances of any one baby being affected by being delivered a week early are “very low”.
The Telegraph ’s headline linking early births with autism is particularly misleading. The term ‘special educational needs’ is applied to several types of disorders in addition to autism, but the study did not look at what specific type of learning difficulties the children had. As such, the results do not support a claim that autism was higher among children delivered early compared with those born at 40 weeks. Autism is a neurodevelopmental disorder that many experts think also has a genetic basis.
What kind of research was this?
Most pregnancies last around 40 weeks. The researchers point out that infants delivered preterm (before 37 weeks) are known to be at increased risk of neurodevelopmental problems including impaired intelligence and school performance, with the risk being highest among the most premature. However, they say there is little information on whether an increased risk also exists among babies born slightly early (37-39 weeks). This is an important issue since these ‘early term’ births are on the increase, and many of these babies are delivered by planned (elective) caesareans.
This was a population-based, retrospective cohort study of 407,503 schoolchildren, which aimed to investigate the risk of special educational needs (SEN) at school age, according to gestational age at delivery. In this type of study, researchers examine the records of a group of people to find out how certain factors (in this case, the week of delivery) might affect their health. A retrospective study, in which researchers look back on past events, is considered less reliable than a prospective study, in which researchers select groups of people and then follow them over time, often for a period of several years. Neither type of study on its own can prove that one event (in this case an early term birth) can cause another (developing SEN), although they can indicate that an association exists between two factors.
In this study, researchers looked at the school census data on 407,503 school-aged children in 19 Scottish local authority areas, which recorded details of any SEN the children had. They linked these data with routine birth data on the same children, held in the Scottish Morbidity Record.
What did the research involve?
The researchers used data from the 2005 school census, provided by 19 out of 32 Scottish local authorities. These authorities covered a total population of 3.8 million, equivalent to 74% of the Scottish population. These authorities’ school census data, along with data provided by head teachers, provided complete information on 362,688 children aged up to 19, including details of any special educational needs (SEN) they had. Special educational needs are defined as learning disabilities such as dyslexia, ADHD and autism, or any physical disabilities that affect learning, such as hearing and vision problems.
These data were then linked to the children’s birth data from the Scottish Morbidity Record, which collects detailed information on all women discharged from Scottish maternity hospitals, including gestational age at birth. Certain cases were excluded because they could easily have biased results, for example, where birth weight was recorded as less than 400g or more than 5,000g, or delivery was before 24 weeks or after 43 weeks.
The two sets of data were then combined and analysed using standard statistical techniques in order to find out if associations existed between week of delivery and special educational needs at school age. In this analysis, the researchers also took into account and adjusted for other factors that might have influenced the results, such as maternal age and height, marital status and birth weight.
What were the basic results?
The study found that the earlier a baby was delivered, the greater the risk of them later having special educational needs. Here are the main findings:
- Overall, 17,784 (4.9%) children out of more than 360,000 with complete data were recorded as having SEN.
- Special educational needs were recorded among 1,565 (8.4%) of those born preterm (before 37 weeks) and 16,219 (4.7%) of those born at term (37-40 weeks).
- Babies born at 37-39 weeks were 16% more likely to have SEN, compared with children born at 40 weeks [adjusted odds ratio (OR) 1.16, 95% confidence interval (CI) 1.12–1.20].
- Babies born at 39 weeks were 9% more likely to have SEN than those born at 40 weeks (a statistically significant difference).
- Babies born at 33-36, 28-32 and 24-27 weeks were 1.53, 2.66 and 6.92 times more likely to have SEN as babies born at 40 weeks (also statistically significant).
- While the risk of developing SEN was higher among preterm babies (those born before 37 weeks), compared with early babies, there were many more babies born early than born preterm. This meant early births accounted for 5.5% of cases of SEN compared with preterm deliveries, which accounted for only 3.6% of cases.
- The risk of SEN at school age was higher in babies born after 41 weeks compared with babies born at 40 weeks.
How did the researchers interpret the results?
The researchers conclude that gestational age at delivery strongly affects a child’s risk of having SEN later in life and that the earlier the delivery before 40 weeks, the greater the risk. They believe the tendency for previous research to analyse preterm babies (before 37 weeks) against ‘term’ babies (37-40 weeks) has meant that the risk for early term babies has gone undetected.
Because early term deliveries are more common (in this study, nearly 40% of the children were delivered between 37 and 39 weeks), they are responsible for a greater number of cases of SEN than preterm deliveries. These findings, say the researchers, have important implications for the timing of elective delivery which, ideally, should be delayed until 40 weeks.
This study is the largest of its kind to look at gestational age across the whole spectrum and any associated risk of a child having special educational needs at school age. It has a number of strengths that increase the reliability of the associations seen:
- It used a national population register and, therefore, avoided any bias due to selection of the study group.
- The data it used were obtained from reputable, national sources.
- The included children were from all types of school and, therefore, represented the population as a whole.
- The researchers took into account a wide range of other factors that might have influenced children’s risk of having SEN. These included adjusting for birth weight, known to be a risk factor for neurodevelopmental problems.
Overall, these robust results have raised an important public health issue, since deliveries at 37-39 weeks seem to be on the rise and a proportion of these will be planned early deliveries, for example elective caesareans or induced labour.
However, this type of retrospective, observational study cannot prove that one factor causes another, i.e. it cannot confirm that an early delivery actually causes cases of SEN. Equally, though, the clear increase in risk with increasing prematurity, known as a ‘dose response relationship’, is evidence supporting a potential cause-and-effect relationship.
Another potential limitation is that the average age of the children followed was 12 years, so although it took into account possible confounding factors at birth, the study could not take into account what happened to children in the years between birth and school age. This means that other factors, such as accidents or environmental problems, could have contributed to the risk of SEN. Also, the cause of an early birth may contribute to the risk, meaning that if a baby is already ill, it may need to be delivered early.
While the results suggest an association between earlier birth and greater risk of SEN, it also needs to be emphasised that, for individual mothers who deliver a baby between 37 and 39 weeks, the risk of neurodevelopmental disorder remains very low. Among the babies in this study delivered at 40 weeks, the risk of SEN was equated to approximately 44 cases per 1,000 births. This means that the 9% increase in risk for babies born at 39 weeks would only amount to about an extra three babies in every thousand compared with those born at 40 weeks.
There are also many factors involved in deciding when an elective caesarean delivery should be undertaken, with some experts pointing out that waiting until 40 weeks also carries some risks.