EV-TF activitywith COVID-19 severity and TEEs. (A) Plasma levels of EV-TF activity in patients with moderate (n = 73) or severe (n = 38) COVID-19 and in the presence (TEE-positive, n = 11) or absence (TEE-negative, n = 100) of TEEs. (B) Receiver operating characteristic curve for TEE prediction using EV-TF (red) area under the curve (AUC), 0.851. The Youden index was determined for EV-TF of 78.3 fM with a sensitivity of 0.818 and a specificity of 0.710. (C) Timing of TEE with an EV-TF threshold of 78.3fM, which is higher than 78.3 fM in red and lower than 78.3 fM in blue. (D) Logistic models (Cox regression): (1) unadjusted (EV-TF level only); (2-7) EV-TF adjusted for age, body mass index (BMI), neutrophil level, aspartate aminotranferase (AST) level, LDH level, and IL-6 level; and (8) adjusted with the combination of all parameters. (E-G) Plasma levels of EV-TF, EV-fib (fibrinolytic activity of EVs), and PAI-1 activity in patients with severe COVID-19 compared with patients with septic shock. The horizontal line and the error bars represent median and interquatile range, respectively. The green areas represent the range of values found in healthy donors. ** P < .01; ***P < .001. CI, confidence interval; HR, hazard ratio; mDO, milliOpticalDensity; TEE, thromboembolic event.