Background: Coronavirus disease 2019 (COVID-19) was first detected as a form of atypical pneumonia. COVID-19 is a highly contagious virus, and some patients may experience acute respiratory distress syndrome (ARDS) and acute respiratory failure leading to death. We aim to evaluate the clinical, imaging, and laboratory parameters according to survival time to predict mortality in fatal COVID-19 patients.
Methods: Fatal 350 and survived 150 COVID-19 patients were included in the study. Fatal patients were divided into three groups according to the median value of the survival days. Demographic characteristics and in-hospital complications were obtained from medical databases.
Results: Of the non-survived patients, 30% (104) died within three days, 32% (110) died within 4-10 days, and 39% (136) died within over ten days. Pneumonia on computational tomography (CT), symptom duration before hospital admission (SDBHA), intensive care unit (ICU), hypertension (HT), C-reactive protein (CRP), D-dimer, multi-organ dysfunction syndrome (MODS), cardiac and acute kidney injury, left ventricular ejection fraction (LVEF), right ventricular fractional area change (RV-FAC), and Tocilizumab/Steroid therapy were independent predictors of mortality within three days compared to between 4-10 days and over ten days mortality. A combined diagnosis model was evaluated for the age, CT score, SDBHA, hs-TnI, and D-dimer. The combined model had a higher area under the ROC curve (0.913).
Conclusion: This study showed that age, pneumonia on CT, SDBHA, ICU, HT, CRP, d-dimer, cardiac injury, MODS, acute kidney injury, LVEF, and RV-FAC were independently associated with short-term mortality in non-surviving COVID-19 patients in the Turkish population. Moreover, Tocilizumab/Steroid therapy was a protective and independent predictor of mortality within three days.
Keywords: COVID-19; acute respiratory distress syndrome; echocardiography; mortality.