Autism may have subtypes that are genetically distinct from each other

The age at which children are diagnosed with autism seems to be partly influenced by their genetics, which may also affect how the condition develops.

“This really provides support for the idea that autism is actually potentially multiple conditions,” says Natalie Sauerwald at the Flatiron Institute in New York, who wasn’t involved in the study.

Autism is a neurodevelopmental condition, characterised by difficulties with social interactions, plus restricted behaviours and interests. The World Health Organization estimates that about one in 127 people are autistic.

“The key question that we were interested in was: why is it that there’s some autistic people who are diagnosed later on in life?” says Varun Warrier at the University of Cambridge.

To find out, he and his colleagues compiled data on people who were diagnosed with autism between the ages of 5 and 17. Their carers completed a questionnaire about their social, emotional and behavioural development, to help the team see if these factors were associated with their age at diagnosis.

The sex of a child, as well as their socioeconomic status, have previously been associated with age at diagnosis – with early diagnoses being more common among boys and those with affluent backgrounds – but the team’s analysis showed that these influences are actually quite weak. “None of the factors typically explain greater than 10 per cent of the variance,” says Warrier.

Instead, the team found evidence of autistic children developing along different trajectories. “What we found was that autistic people were grouped into two broad groups,” says Warrier. One group started experiencing difficulties early in life and these stayed largely constant. The other had fewer difficulties in early childhood, but experienced more in late childhood and early adolescence. This explained “anywhere between 10 to 25 per cent of the variance in age of autism diagnosis”, he says.

Further, the researchers found that these differences between the two groups were reflected in the children’s DNA, which had previously been collected. Those who were diagnosed later tended to have a different set of common genetic variants than those diagnosed earlier, which explained 11 per cent of the variation in age of autism diagnoses. The team didn’t have the necessary data to look for rare variants or those that arise spontaneously, rather than being inherited.

Despite these two trajectories, there is no sign of a hard separation between them, says Sauerwald. “You have overlap between the groups, probably because they’re not perfectly separable,” she says. Warrier similarly describes them as existing “on a gradient”.

He emphasises that neither group should be thought of as having milder or more severe autism, which research increasingly suggests takes multiple forms. In July, Sauerwald and her colleagues published a study in which they found evidence of four moderately distinct groups, with differing symptoms, behaviours and underlying genetics.

Warrier’s team also found that the children diagnosed later were more likely to have genetic variants associated with other conditions, including ADHD and post-traumatic stress disorder. In Sauerwald’s study, her team found a similar link between late autism diagnosis and ADHD, which, like autism, is a form of neurodiversity. Why this link exists is unclear, says Sauerwald. “There’s a surprising lack of data on ADHD,” she says, especially about the specific symptoms people experience.

Better understanding the potential subtypes of autism could eventually help improve diagnostic practices and lead to more tailored support for autistic children and their families. “That can only lead to improvements in the quality of life for individuals – when we can better understand their condition, and how to help them, and what it is that they need exactly and what they don’t need,” says Warrier.