Cytokine production profiles can predict COVID-19 severity

Hiroshi Morimoto

COVID-19, SARS-CoV-2, cytokine storms, IL-6, IL-17, IL-16.

Mortality in COVID-19 patients is related to the presence of a “cytokine storm” induced by the virus. Most patients developed mild symptoms, whereas some patients develop severe disease. Predicting the course of disease is necessary to mitigate or prevent COVID-19 disease severity. Carefully monitoring specific cytokines during the management of COVID-19 patients might improve patients’ survival rates and reduce mortality from COVID-19. For example, IL-6 levels in patients with COVID-19 had been considered a relevant parameter in predicting the most severe course of the disease. The purpose of this study is to investigate whether a patient’s cytokine levels would predict the course of disease, and to describe the characteristic differences in cytokine levels between patients with no symptoms and those with severe disease. We applied a probabilistic method, naive Bayes classifier, to RNA-sequencing data extracted from GEO with the accession number GSE178967. We predicted a patient’s disease course, i.e. either deterioration or improvement, and calculated the comprehensive accuracy of our prediction. There were characteristic cytokine level patterns preceding a severe state of disease. Some important cytokines were identified other than IL-6 and IL-17, which are already known as key cytokines associated with a cytokine storm. Our methodology shows that the systematic observation of cytokine levels in patients with COVID-19 can yield important information in predicting the most severe course of disease and thus the need for appropriate and intensive care.

https://dx.doi.org/10.31873/IJEAS.10.2.03