Validation of biomarkers predicting prognostic severity at the onset of COVID-19 after corona-vaccine dissemination: Comparative research from Bangladesh and Japan

Abstract

Abstract This study utilized data on serum cytokines, chemokines, and soluble receptors of 129 Bangladesh COVID19-infected patients within days of disease onset and prior to treatment. The majority of patients in Bangladesh between December 2021 to September 2022 was infected with the Omicron strain, and over 88% of them were vaccinated. Patients were divided into two groups: those who were cured with mild disease and who deteriorated and progressed to moderate and severe disease. This data was used as training data to identify markers with the LASSO method, and a formula containing 15 predictive markers was selected. The predictive markers were then used to predict the final outcome for 194 Japanese patients. The Japanese patients transitioned from the Wuhan to the Omicron strain, with data collected for almost two years from June 2020 to June 2022. Japanese patients’ results showed a good AUC of 0.9219. We observed that the percentage of correct answers for the virus strain in waves 5 and 6 (over 80%) were higher than the strains in waves the 2 to 4. These results are particularly useful as predictive markers of exacerbation to determine whether post-vaccine COVID infection can be handled with medication and home care, or whether hospitalization is required.