Birthweight link to autism not clear

“Babies born weighing less than 4lb (1.8kg) could be more prone to developing autism than children born at normal weight,” BBC News has reported.

This finding comes from a study that found about 5% of infants whose birthweight was less than 2000g (about 4lbs and 6oz) had autistic spectrum disorders (ASDs) at the age of 21. This was higher than previous estimates that suggested that 0.9% of US eight year olds of any birthweight have been diagnosed with some form of ASDs.

The main limitation to this study is that it did not include a control group of children with normal birthweight to compare against those with low birthweight. Instead, it relied on general population estimates to examine the relationship. This makes understanding the issue more complex as the children in this study were all given specific assessments for detecting ASDs that would not be routinely given to children in the general population.

This means we cannot be certain to what extent children with low birthweight truly do have higher rates of ASDs or if the methods used simply detected cases that would go undiagnosed in everyday life. This is supported by the fact that some of the cases identified had not previously been diagnosed by a doctor.

It is also worth noting that a large proportion of those eligible to participate did not complete the study and this may have influenced results. Overall, the findings of this study need to be confirmed by more robust studies with a control group of infants of normal birthweight.

Where did the story come from?

The study was carried out by researchers from the University of Pennsylvania and other research institutes in the US. It was funded by US National Institutes of Health and published in the peer reviewed medical journal Pediatrics.

BBC News generally placed the study into context well, noting that the findings need to be confirmed in other studies and including quotes such as those from Dorothy Bishop, professor of developmental neuropsychology at the University of Oxford. She is quoted as saying, 'the association looks real, but nevertheless, most low birthweight children don't have autism, and most children with autism don't have low birthweight'.

What kind of research was this?

This was a prospective longitudinal study that looked at what proportion of infants with low birthweight went on to develop autistic spectrum disorders (ASDs) in adolescence or early adulthood.

ASDs, including autism and Asperger’s syndrome, and are a group of related disorders that begin in childhood and persist into adulthood. They are diagnosed by the presence of three broad categories of symptoms:

difficulties with social interaction
impaired language development and communication skills
unusual patterns of thought and physical behaviour

In over 90% of cases no underlying medical condition can be found to explain the symptoms of ASD, though causes continue to be investigated.

The researchers say that low birthweight is an established risk factor for cognitive and movement problems, and some studies have suggested that low birthweight may also be a risk factor for ASDs. However, they also point out that most prospective studies examining this possible association have not made firm diagnoses of ASDs using standard diagnostic methods.

This study followed only a group of low birthweight individuals, and then made comparisons with how common autism is in the population as a whole, based on the figures reported in another study. This may give some idea of whether autism is more common in low birthweight infants but there are some limitations. For example, the children in this study were given assessments to specifically test whether they had autism, which means that more cases might be picked up than would be found than in the general population, which is not routinely screened for autism.

Ideally, the study would have included groups of babies with different birthweights all born in the same period, and followed and assessed them in the same way. This could help establish whether the results seen would be genuinely due to increased prevalence or due to increased diagnosis. It would also allow them to take into account any other differences between babies of low and normal birthweights.

What did the research involve?

The researchers enrolled 1,105 low birthweight infants who weighed less than 2000g when they were born. Babies born between October 1, 1984 and July 3, 1989, at three New Jersey hospitals were eligible. Of these infants, 862 (78% of those enrolled) were eligible for follow-up at age 16, and 623 (56%) were screened to identify those that might have ASDs. At age 21, the researchers used standard diagnostic interviews for ASDs to reassess 60% of those who screened positive for autism at age 16, and 24% of those who tested negative. This confirmed which individuals had a diagnosis of ASDs. They used the figures to estimate how common autism was in the entire group of low birthweight babies.

The infants in this study were part of the Neonatal Brain Hemorrhage Study (NBHS), which included all infants admitted to three hospitals in New Jersey, which cared for 85% of low birthweight babies born in the area. The children were assessed at ages 2, 6, 9, 16, and 21 years. At age 16, this included questionnaires that the parents completed about autism symptoms and social communication. The questionnaires asked parents if their child had ever been diagnosed with ASDs by a healthcare professional.

The ASDs the researchers tested for included autism, Asperger’s syndrome, or a pervasive developmental disorder (not otherwise speciﬁed). Those scoring over set thresholds on the questionnaires or those with a professional ASD diagnosis were considered to be ‘screen positives’, and assessed again for a ASDs diagnosis at age 21.

The researchers also tested a proportion of adolescents who had screened negative at 16 to identify whether the initial screening had missed any cases. The diagnostic interviews at age 21 were conducted with the parents and with the young adult themselves, and were performed by researchers who did not know whether the participants had screened positive or negative for ASDs at age 16.

What were the basic results?

In the first ASDs screen at age 16 years, 117 low birthweight adolescents (18.8% of those tested) screened positive for ASDs. Of these 117 adolescents, 47 (40.2%) were lost to follow-up or did not complete the ASD questionnaires at age 21. Of the 70 who were assessed at age 21 years, 11 (15.7%) were confirmed as having ASD at age 21.

In the first ASD screen at age 16 years, 506 (81.2%) low birthweight adolescents had screened negative for ASDs. Of these 506 adolescents, 119 (23.5%) were selected for assessment at age 21. Of these 119 screen negatives, three (2.5%) were found to have ASD at their later assessment.

Most of the participants identified with ASDs (9 out of 14) were reported to have relatively high levels of functioning, spoken language, and with IQs of 70 or over.

Based on these figures, and the proportion of screen positives and screen negatives at age 16, the researchers calculated that about 5% of the entire low birthweight cohort assessed at age 16 had ASD. Just over half of these young adults (8 out of 14) had not been diagnosed prior to this study.

There were some differences between the individuals who could be followed up and those who could not. For example, those who were not followed up at age 21 were more likely to have had suboptimal neurodevelopmental outcomes at age 16 (for example, cognitive or movement disabilities).

How did the researchers interpret the results?

The researchers conclude that ASDs in this group of low birthweight individuals was 5%. They say that this was higher than the prevalence of 0.9% reported by the US Centers for Disease Control and Prevention for eight year olds in the general US population (all birthweights) in 2006.

Conclusion

This study has suggested that about 5% of children of low birthweight (<2000g) in the US may go on to develop autistic spectrum disorders (ASDs). This is higher than previous estimates for the general population of children in the US (estimated at 0.9% among eight year olds). When considering these results there are both strengths, such as the prospective nature of the study, and limitations that must be taken into account:

The children in this study were specifically assessed to see if they had autism, which means that more cases might be picked up than would be found than in the general population, who are not all assessed for autism. Ultimately, this raises the question of whether the results reflect greater prevalence among underweight babies or greater rates of diagnosis. Ideally, the study would have included a group of babies of different birthweights all born in the same period, and followed them all up and assessed them in the same way. This would also allow them to take into account any other differences between low birthweight and normal birthweight babies that could influence rates of ASDs.
The general population prevalence figures were based on children aged eight, and these figures may differ from those found among adolescents and young adults such as those in this study.
A high proportion (40%) of those assessed at age 16 could not be assessed again at age 21, and this could have influenced the results.
Not all participants received all parts of the screening and diagnostic assessments.

The results of this study need to be confirmed by more robust studies with a control group of infants with normal birthweight. It is also worth bearing in mind the quote from Dorothy Bishop, professor of developmental neuropsychology at the University of Oxford, in the BBC News: 'The association looks real, but nevertheless, most low birthweight children don't have autism, and most children with autism don't have low birthweight.'