Autism can be diagnosed using a stool sample, scientists say
Scientists have raised hopes that a cheap and simple test for autism is possible after discovering consistent differences between the microbes in the guts of autistic people and those of people without autism.
The findings suggest that a routine stool test could help doctors detect autism early, meaning people could be diagnosed and hopefully supported much sooner than the lengthy procedure currently used in clinics.
“It usually takes three to four years to get a confirmed diagnosis for suspected autism, with most children being diagnosed by the age of six,” said Prof Qi Su from the Chinese University of Hong Kong. “Our microbiome biomarker panel performs very well in children under the age of four, which may help to make an early diagnosis.”
Autism rates have risen dramatically in recent decades, largely due to greater awareness and a broadening of the criteria used to diagnose the condition. In the UK and many other Western countries, it is now thought that around one in 100 people are on the autism spectrum.
Studies of twins suggest that 60-90% of autism is due to genetics, but other factors contribute, such as older parents, complications at birth, and exposure to air pollution or certain pesticides during pregnancy. Signs of autism range from children who don’t respond to their name and avoid eye contact, to adults who have a hard time understanding what others are thinking and become anxious when their daily routines are disrupted.
Scientists have known for some time that autistic people often have fewer diverse bacteria in their digestive tracts. Whether this is somehow due to autism or whether it actually contributes to the condition is still a matter of debate.
To dig deeper into the puzzle, Su and his colleagues analyzed stool samples from 1,627 children ages 1 to 13, some of whom were autistic. They checked the samples to see what bacteria were present, and did the same for viruses, fungi and other microbes called archaea.
Write in Nature Microbiologythe researchers describe how gut microbes differed significantly between children with and without autism. In total, 51 species of bacteria, 18 viruses, 14 archaea, seven fungi, and a dozen metabolic pathways were altered in autistic children. Armed with machine learning, a form of artificial intelligence, the scientists were able to identify the autistic children with 82 percent accuracy, based on 31 microbes and biological functions in the digestive system.
The study also revealed other changes: several metabolic processes involved in energy and neurological development were found to be disrupted in the autistic children.
“While genetic factors play a substantial role in autism, the microbiome may act as a contributing factor by modulating immune responses, neurotransmitter production, and metabolic pathways,” Su said. “This does not necessarily imply causality, but suggests that the microbiome may influence the severity or expression of autism spectrum symptoms.”
If the researchers’ thinking is correct and disrupting the microbiome does indeed influence the severity of autism, it raises the prospect of personalized interventions using diet or live bacteria, also called probiotics, to create a more diverse microbiome in people diagnosed with autism.
“Ultimately, this broad scope increases the potential to develop more effective, noninvasive diagnostic tools and therapeutic strategies for autism,” Su said. The team is now conducting a clinical trial to investigate whether stool samples can help identify autistic children as young as one year old.
Dr Dominic Farsi of King’s College London said the results could have “great potential” in diagnostic practice, but added that more research was needed to confirm the findings. “Nevertheless, these results could represent a major step towards improving diagnostic methods for autism spectrum disorder,” he said.
Dr Elizabeth Lund, an independent nutrition and gastrointestinal health consultant, said: “The idea that stool sample analysis could aid diagnosis is very exciting as there is currently a huge backlog of children and adults waiting for assessment. The current process is very lengthy and there is a shortage of clinicians such as psychologists and psychiatrists who are trained to make an accurate diagnosis.
“Clearly the research needs to be replicated by other groups and in other populations around the world, but the approach could provide a new and more automated route to diagnosis in the longer term.”