AI, blood biomarker spots long Covid patients with 96% accuracy

NEW YORK: Scientists have, for the first time, identified specific blood biomarkers that can accurately identify patients with long Covid.

In the research, published in the journal Nature, they used an artificial intelligence based algorithm and showed that long Covid patients have clear differences in immune and hormone function from patients without the condition.

"These findings are important -- they can inform more sensitive testing for long Covid patients and personalised treatments for long Covid that have, until now, not had a proven scientific rationale," said principal investigator David Putrino, Professor at Icahn Mount Sinai.

"This work is so exciting because it is one of the first to show us clear, measurable differences in blood biomarkers of people with long Covid compared with people who recovered fully from an acute infection and a group of people who have never been infected with SARS-CoV-2 (the virus that causes Covid-19).

"This is a decisive step forward in the development of valid and reliable blood testing protocols for long Covid," Putrino said.

The team analysed a total of 271 patients between January 2021 and June 2022 using machine learning-based algorithm.

They were divided into three groups: those with no prior SARS-CoV-2 infection; those who had fully recovered from a clinically confirmed case of Covid-19; those with active long Covid symptoms for at least four months or more after confirmed Covid-19 infection (median time of long-term symptoms was 12 months since the acute infection).

Overall, the algorithm was able to differentiate between people with and without long Covid with 96 per cent accuracy and detect the condition based on distinctive features detected in the blood of participants in the long Covid group, the researchers said.

Some of the most pronounced differences between the long Covid group and the two control groups were related to immune and hormonal dysfunction. This was characterised by biomarkers indicating abnormal T cell activity, reactivation of multiple latent viruses (including the Epstein-Barr virus and other herpesviruses) and significant reductions in cortisol levels.

"These findings show us that people with long Covid are living with a disease process that is observable using the blood testing protocols laid out in the study, but also varies from patient to patient depending on their specific medical history," Dr. Putrino said.

"This means that physicians must listen to their patients and perform a wide variety of physiological and lab tests, while adopting a highly personalised approach to the medical management of long Covid.

"There is no 'silver bullet' for treating long Covid, because it is an illness that infiltrates complex systems such as the immune and hormonal regulation. Complex illnesses require complex treatment solutions and we need more rapid research to better understand long Covid and discover new and promising therapies," Dr. Putrino said.