ABSTRACT
Background: Several studies suggest an increased incidence of thrombosis in COVID-19 patients. However, evidence on how to prevent and even treat it is scarce. The aim of this study was to compare the cumulative incidence of venous thromboembolism (VTE) of two different methods for lower extremity deep vein thrombosis (LE-DVT) diagnosis: systematic vs. clinically guided complete compression venous ultrasonography (CCUS). We conducted a monocentric, prospective, open-label, non-randomized study. All consecutive patients admitted in three intensive care units (ICUs) of University Hospital of Toulouse for COVID-19 pneumonia were included: one performed systematic screening for LE-DVT, the others did not. The primary outcome was the 21-day cumulative incidence of VTE. The secondary end points were the 21-day cumulative incidences of major bleeding and death. Results: Among the 78 patients included, 27 (34.6%) underwent systematic screening for DVT 7 ± 2 days after ICU admission. Thirty-two patients (41.0%) were diagnosed with VTE, with a 21-day cumulative incidence of 42.3% (95% CI, 31.4-55.2), without difference between screened and non-screened patients (hazard ratio 1.45, 95% CI, 0.72-2.93). In the screened group, the frequency of isolated DVT was higher (25.9 vs. 5.9%, p-value = 0.027), but the frequency of pulmonary embolism was not reduced (25.9 vs. 29.4%, p-value = 0.745). The 21-day cumulative incidences of major bleeding and death were 9.6% (95% CI, 4.7-19.2) and 10.3% (95% CI, 5.0-20.8), respectively, without difference between the two groups. Conclusions: A systematic screening for DVT in patients hospitalized in ICU was not associated with a higher diagnosis of VTE or a reduced diagnosis of PE.