Is It Possible To Predict Mortality Using Initial Data Of Adult Patients Hospitalized with COVID-19? A Mortality Prediction Model in the Early Phase of COVID-19

Abstract

Objective: In this study, we aimed to determine the factors that contribute to the early determination of mortality risk in patients hospitalized with COVID-19. Methods: We included 941 adult inpatients (474 male [50.4%], mean age, 53.5 +/- 17.0. The patients were divided into two groups: the discharge group and the death group. Epidemiological data, medical history, underlying comorbidities, laboratory findings, chest computed tomographic scans, real-time reverse transcription polymerase chain reaction detection results, and survival data were obtained with retrospective recordings on admission and follow-up. The statistical relationship between survival data and parameters was analyzed. A mathematical model was created from the data of both groups. Results: While 863 patients survived, 78 were non-survivors. During the study period, the preliminary case fatality rate of the inpatients was 8.3%. The mean age of the non-survivors was 71.7 +/- 11.2 SD (P<0.001). Laboratory findings showed that mortality was high in those with high D-dimer, sodium, lactate dehydrogenase (LDH), troponin, creatine kinase-myocardial band (CK-MB), ferritin, blood lactate, activated partial thromboplastin time, and high blood glucose levels (P<0.05). Furthermore, mortality was high in patients with low albumin, lymphocyte, and platelet levels (P<0.05). The logistic regression model showed that advanced age, hypertension, high D-Dimer (>1000 ng/ml), high C-reactive protein (CRP), CK-MB, and LDH, and low lymphocyte count were associated with poor prognosis. Conclusions: According to week 1 data of patients with COVID-19, advanced age, hypertension, D-Dimer, CRP, CK-MB, high LDH, and low lymphocyte were associated with poor prognosis. We believe that this model will be useful in predicting patient mortality.