Hematological and coagulation parameters as predictors of death by Coronavirus disease in hospitalized patients: a Brazilian follow-up study

Abstract

Abstract This study aimed to evaluate the hematological and coagulation parameters according to the clinical outcomes of coronavirus disease (COVID-19). We analyzed the hematological and coagulation parameters of hospitalized patients with COVID-19 at admission, and two and three weeks during hospitalization. To assess the performance of these parameters in predicting poor outcomes, receiver operating characteristic (ROC) curves were created. We studied 128 patients with COVID-19 (59.2±17.7 years, 56% male). Non-survivors (n=54, 42%) presented significant alterations in hematological and coagulation parameters at admission, such as increased in white blood cells (WBC), neutrophil, and band cell counts, as well as elevated prothrombin time (PT), activated partial thromboplastin time, and D-dimer levels. During follow-up, the same group presented a gradual increase in D-dimer and PT levels, accompanied by a reduction in PT activity, hemoglobin, and red blood cell count (RBC). ROC curves showed that WBC, neutrophil, and band cell counts presented the best area under the curve (AUC) values with sensitivity and specificity of >70%; however, a logistic regression model combining all the parameters, except for RBC, presented an AUC of 0.89, sensitivity of 84.84%, and specificity of 77.41%. Our study shows that significant alterations in hematological and coagulation tests at admission could be useful predictors of disease severity and mortality in COVID-19.

Keywords

• COVID-19
• Hematology
• Coagulation
• Predictors
• Death