Empiric Anticoagulation Therapy in Hospitalized COVID-19 Patients: An Evaluation of Bleeding Risk Scores Performances in Predicting Bleeding Events

Abstract

Currently, there is no standardized consensus on anticoagulation (AC) among patients with coronavirus disease (COVID-19), which has an overwhelming bleeding risk. We aimed to compare the patterns of AC in COVID-19 patients and compare two validated risk scores in predicting bleeding events. A retrospective review of medical records was conducted for COVID-19 patients who received empiric anticoagulation therapy. The primary outcomes included bleeding events, survival, and mechanical ventilation needs. We applied the HAS-BLED and ORBIT bleeding risk scores to assess the predictive accuracy, using c-statistics and the receiver operating curve (ROC) method. Of the included patients (n = 921), with a mean age of 58.1 ± 13.2, 51.6% received therapeutic AC and 48.4% received a prophylactic AC dose. Significantly higher values of d-dimer and C-reactive protein (CRP) among the therapeutic AC users (p < 0.001) were noted with a significantly prolonged duration of hospital stay and mechanical ventilation (p < 0.001 and p = 0.011, respectively). The mean value of the HAS-BLED and ORBIT scores were 2.53 ± 0.93 and 2.26 ± 1.29, respectively. The difference between the two tested scores for major bleeding and clinically relevant non-major bleeding was significant (p = 0.026 and 0.036, respectively) with modest bleeding predictive performances. The therapeutic AC was associated with an increased risk of bleeding. HAS-BLED showed greater accuracy than ORBIT in bleeding risk predictability.