New autism blood test likely to join ranks of also-rans

Post Reply New Topic

not only does it promise a cheaper, faster, simpler way to diagnose a mysterious condition, it may even offer the possibility of earlier diagnosis.

The test was developed by taking blood samples from 159 children under 10, about half of whom had autism. They took 24 different measurements relating to a range of metabolic pathways or other biochemical markers previous research deemed relevant to autism.

Of these 24, the researchers found five that could separate a number of children with autism from another “neurotypical”, or non-autistic group. Over 96 per cent of children in each group got a result that matched the standard behavioural diagnosis.

That sounds like an impressive result. However, there is a problem with any potential test: because autism is relatively rare – affecting about 1 per cent of the population – the number of children with autism diagnosed correctly would be outnumbered by the neurotypical children also given that diagnosis. So a positive result would probably be wrong more often than not.

That doesn’t mean the test has no value. It could still be used to find out which children should get behavioural testing. But there’s a deeper issue: we don’t yet know if this is a test for autism or instead a more general test for brain differences.

The five blood compounds in the test aren’t specific to the changes that seem to link to autism. Some are in pathways involved in helping cells resist damage from inflammatory chemicals or toxins. Others are involved in regulating how genes are turned on or off – which genes might be affected here are unclear.