ABSTRACT
Background: Recovery from coronavirus disease 19 (COVID-19) is a gradual process that depends on the disease severity. Immunologic changes that precede and relate clinical symptoms may predict the course of COVID-19 and final outcome. Our goal was to determine prognostic markers of COVID-19 improvement. Method(s): The study included hospitalized patients from the ages 31-72 with moderate to severe COVID-19. All biomarkers were assessed at three checkpoints starting from the first day of hospitalization (day 0), continuing on day 8, and between 40-50 day. Luminex xMAP technology and the Bio-Plex Pro Human Cytokine 17-plex assay was used for quantitative evaluation cytokines and chemokines in peripheral blood of COVID-19 patients. The comparative study was done in combination with clinical data. Univariate and multivariate analyses of data were delivered. Finally, a fuzzy logic model for decision support was proposed and validated for explored data. Result(s): Macrophage inflammatory protein-1beta (MIP1b) was inversely related to COVID-19 evolution. MIP1b significantly higher on day 8 compared to day 0 (p < 0.0001) correlated with clinical improvement and predicted a successful course of the disease. It was also associated with the significant increase in TNF-alpha (p = 0.03), and decrease in IL-10 (p < 0.0001), and IL-6 (p = 0.01). The increase in MIP1b on day 8 correlated positively with eosinophil and lymphocytes counts and negatively with inflammatory mediators (ferritin, procalcitonin, fibrinogen, CRP). Moderately positive correlation between MIP-1b and TNF-alpha was noted, in parallel. Tested the statistical and machine learning predictors exhibited sensitivity to MIP1b input, improving the ROC curve compared to the classification models trained without MIP1b. Conclusion(s): This finding next to already known indicators such as IL-6, eosinophil and lymphocytes counts, highlight a role of MIP1b as a marker of good prognosis in COVID-19 and provide a novel insight into this as a potential diagnostic and therapeutic target.