“Women who delay pregnancy are more likely to have a child with autism,” the Daily Mail reported. It said researchers have found that a woman of 40 has a 50% higher risk than a woman in her late 20s.
This research followed nearly 5 million children from birth and compared the parental characteristics of those who developed autism with those who did not. It found that older mothers were associated with an increased risk of the child later developing autism.
This was generally well-conducted research, but it only considered a few of the many possible factors that might affect the risk of autism. The cause of autism is not known, but genetics, brain development, allergies, immunity and the environment have all been suggested as possibilities.
Older women who want to have children should not be overly concerned by these findings. Their risk of having a child with autism remains small. Overall, only about 0.2% of the children in this study developed autism. A systematic review of these results and other similar studies may be able to determine whether the evidence supports a link between parental age and autism risk.
The research was carried out in the US by Janie F Shelton and colleagues from the University of California. The study was funded by grants from the National Institute of Environmental Health Sciences, the US Environmental Protection Agency, and the UC Davis School of Medicine and Office of Graduate Studies. The paper was published in the peer-reviewed medical journal Autism Research .
The news stories have generally reflected the findings of this paper accurately.
Previous studies investigating if parental age affects the risk of having a child with autism have given inconclusive results. This study used a birth cohort (from birth onwards), with data available on the parents’ age and education. Using people from this cohort, a case-control study was carried out comparing the children who had developed autism with the rest of the birth cohort (the controls).
Cohort studies are the most appropriate type of study for observing whether a particular exposure affects the risk of a later outcome. This study had strengths in that it followed a large group of children (4,947,935) over 10 years, and that the exposure (mother’s age at child’s birth) definitely preceded the development of autism - essential for studies aiming to establish causation.
However, to further strengthen the validity of these results, steps need to be taken to ensure that other possible confounding factors have been taken into account. The difficulty with autism is that its causes are not known, so it is difficult to take into account all the possible factors that could affect risk.
The researchers obtained the records for all births in California between January 1990 and December 1999. Cases of autism were identified from this cohort using the records from routine examinations called the Early Start Report (ESR) for children under three, and the Client Development and Evaluation Report (CDER) for children over three.
A diagnosis of autism was defined as either a checkmark for autism under Developmental Disabilities on the ESR, or an autism level of one (Full Syndrome Autism) on any CDER record, or an ICD code (a standard diagnostic code) for autistic disorder. Diagnosis data were available through to the year 2006. After excluding children from multiple births and those with missing data on parental age and education, there were 12,159 cases and 4,935,776 controls.
The researchers used this data to construct models of the relationship between parental age and autism risk, which were adjusted for the potential confounders of the parents’ race or ethnicity, their number of previous pregnancies and births, year of birth, insurance type and sum of parental education (as a proxy measure of socioeconomic status). Parental age at birth was split into categories of under 25; 25 to 29 (which acted as the reference group for other ages); 30 to 34; 35 to 39; and over 40.
A higher proportion of the children with autism were male. Compared to controls, autism cases were more likely to have older parents, and to be of either non-Hispanic, white or Asian ethnicity. The median (average) age of mothers at the time of delivery was 30 for cases and 27 for controls, while for fathers these figures were 32 and 29, respectively.
Advancing maternal age was found to be associated with an increased risk of autism. When other possible confounders had been adjusted for, a mother of 40 or over at the time of birth was 51% more at risk of having a child with autism compared to a mother aged 25 to 29 (odds ratio [OR] 1.51, 95% confidence interval [CI] 1.35 to 1.70), and 77% more at risk compared to a mother aged less than 25 (OR 1.77, 95% CI 1.56 to 2.00).
For a mother, the risk of having a child with autism was not influenced by the father’s age in any way. The relationship with father’s age was not so clear. It appeared that fathers of 40 or over only had an increased risk of having a child with autism if the mother was under the age of 30 (OR 1.59, 95% CI 1.37 to 1.85). If the mother was over the age of 30, the risk from the father being 40 or over was only of borderline significance (OR 1.13, 95% CI 1.01 to 1.27).
The researchers conclude that a woman’s risk of having a child who later develops autism increases throughout her reproductive years. However, a man’s risk of having a child with autism seems less influenced by his age and more by his partner’s.
This appears to be generally well-conducted research. It followed a large cohort of 4,947,935 children from birth to 6-16 years and compared the parental characteristics of 12,159 children who developed autism with the parents of children who did not develop autism. The study found older maternal age at birth increases the child’s risk of autism.
There are a few points to bear in mind with this study. The main one is that it took into account only a few of the factors that potentially affect the risk of autism (mainly proxy measures of socioeconomic status).
The causes of autism are not known. Genetics, brain development, allergies, immunity and the environment have been suggested, but so far are only speculative.
Several minor limitations include the possibility that the children with autism were misdiagnosed or miscoded in the database, and that it may not be possible to directly apply the results outside of California, as different social and cultural environments may affect risk.
Older women who are planning to have a child should not be overly concerned by these findings. Their risk of having a child with autism remains small. Overall only about 0.2% of the children in this study developed autism.
The authors report that other studies looking at the same question have had varied results. A systematic review looking at all such studies together may be able to determine why this is the case and whether the evidence as a whole supports a link between parental age and autism risk.