About one-fifth of patients with coronavirus disease 2019 (COVID-19) have clinically severe or life-threatening infections requiring interventions such as immediate oxygen therapy or mechanical ventilation. Knowing early on which patients are likely to develop severe disease could help save lives.
To that end, several research teams are working to discover telltale blood biomarkers. In a study recently published in Cell, researchers in China found 93 proteins and 204 metabolites whose levels correlated with severe COVID-19. The scientists analyzed sera from 53 healthy people and 46 patients with severe and nonsevere COVID-19 to find the molecular markers.
They then trained a machine learning model to stratify disease severity using 29 of these serum factors. The model correctly classified 29 of 31 patients in the training cohort, for an overall accuracy of 93.5%, and 23 of 29 patients in 2 independent test cohorts.
Larger studies of patient samples collected at more time points will be needed to develop a clinical test that predicts severe cases before they develop. Still, the study provides “some of the first evidence that such a test might be possible,” National Institutes of Health Director Francis Collins, MD, PhD, wrote in a recent blog post.
Meanwhile, European researchers have designed a high-throughput platform to analyze serum and plasma proteins from COVID-19 samples. The system identified 27 potential protein biomarkers that are expressed differently in hospitalized patients depending on their case severity. Another team, led by New York University researchers, has developed a point-of-care mobile app that provides a COVID-19 severity score based on patients’ biomarker measurements and clinical risk factors.
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.