Selected Predictors of COVID-19 Mortality in the Hospitalised Patient Population in a Single-Centre Study in Poland

Abstract

Background: The correct analysis of COVID-19 predictors could substantially improve the clinical decision-making process and enable emergency department patients at higher mortality risk to be identified. Methods: We retrospectively explored the relationship between some demographic and clinical factors, such as age and sex, as well as the levels of ten selected factors, namely, CRP, D-dimer, ferritin, LDH, RDW-CV, RDW-SD, procalcitonin, blood oxygen saturation, lymphocytes, and leukocytes, and COVID-19 mortality risk in 150 adult patients diagnosed with COVID-19 at Provincial Specialist Hospital in Zgierz, Poland (this hospital was transformed, in March 2020, into a hospital admitting COVID-19 cases only). All blood samples for testing were collected in the emergency room before admission. The length of stay in the intensive care unit and length of hospitalisation were also analysed. Results: The only factor that was not significantly related to mortality was the length of stay in the intensive care unit. The odds of dying were significantly lower in males, patients with a longer hospital stay, patients with higher lymphocyte levels, and patients with higher blood oxygen saturation, while the chances of dying were significantly higher in older patients; patients with higher RDW-CV and RDW-SD levels; and patients with higher levels of leukocytes, CRP, ferritin, procalcitonin, LDH, and D-dimers. Conclusions: Six potential predictors of mortality were included in the final model: age, RDW-CV, procalcitonin, and D-dimers level; blood oxygen saturation; and length of hospitalisation. The results obtained from this study suggest that a final predictive model with high accuracy in mortality prediction (over 90%) was successfully built. The suggested model could be used for therapy prioritization.