Researchers turn to AI for help predicting COVID severity in children

While children infected with COVID-19 have tended to have milder symptoms, researchers note some have developed severe complications such as respiratory failure or heart inflammation.
Jeff Rowe

Throughout the COVID-19 pandemic, children infected by the virus have generally suffered milder symptoms than adults, but will that be the case if and when new variants emerge?

That’s the question researchers are working to answer in an ongoing study that’s examining the relationship between cytokines in saliva and COVID-19 infection. Cytokines are proteins found in blood and saliva that could produce a response to a COVID-19 infection.

According to a recent update on the team’s progress, in a preliminary analysis of saliva samples from 150 children the researchers discovered that levels of two cytokines were higher in those with severe COVID-19 compared to those without.

“Using saliva to predict severity of the infection is non-invasive and painless,” explained study author Usha Sethuraman, MD, in the statement. “If proven to be effective saliva may be a game-changer in children in whom obtaining blood is both difficult and distressing. Additionally, early recognition of the severity of COVID-19 can help clinicians institute timely and appropriate treatment which may help improve outcomes.”

The goal of the study is to be able to identify children at risk for severe disease by integrating biomarkers such as cytokines and social determinants of health using AI. Studies in adults have demonstrated that certain cytokines are elevated in the blood of patients with COVID-19, and predictive analytics could help determine the infection severity.  The researchers are also in the process of gathering saliva samples from 400 children ages 18 and young who have tested positive for COVID-19.

According to the statement, “In addition to finding that levels of two cytokines (MIG and CXCL-10) in the preliminary analysis were higher in those with severe COVID-19 compared to those without severe infection, dozens of microRNA levels were found to be altered, with the majority of them being significantly lower in the saliva of children with severe infection.”

The patient data is coming from Children’s Hospital of Michigan and UPMC Children’s Hospital of Pittsburgh, while the saliva sample analysis is being performed at Penn State College of Medicine.  In addition, Wayne State University is creating a model using AI.

Ongoing analysis will pursue the validation of the results and confirm the importance of saliva cytokines and microRNAs, combined with social factors, including where a child lives.

The study is supported by a grant from the Eunice Kennedy Shriver National Institute of Child Health and Human Development.

Photo by portishead1/Getty Images