Risk factors associated with deep vein thrombosis in COVID‐19 patients

Abstract  


Dear Editor,
The ongoing coronavirus disease 2019 (COVID-19) pandemic has become the greatest threat to the public health of this generation. Patients infected with severe acute respiratory syndrome coronavirus 2 have a high chance of developing acute respiratory distress syndrome (ARDS) and the mortality rate is high. 1,2 Risk factors associated with mortality include age, neutrophilia, organ abnormality, and coagulation dysfunction (e.g., elevated D-dimer). 1 Consistent with previous findings that viral infections are able to induce thrombosis in human and a broad range of animal hosts, 3 elevated coagulation function in COVID-19 patients might enhance the risk of venous thromboembolic events, leading to multifactorial thrombotic disorder such as deep venous thrombosis (DVT) and pulmonary embolism (PE), occasionally as thrombosis of the hepatic, portal, or splanchnic veins. However, the risk factors associated with thrombus formation remained largely indeterminate. Understanding the underlying causes of thrombosis can have important implications for choice and dosing of antithrombotic interventions to enhance diagnostic care and patient management.
In this single-center study, we retrospectively analyzed the bedside ultrasound findings in COVID-19 patients with advanced disease admitted to Union Hospital of Tongji Medical College (Wuhan, China) from January 29 to April 11, 2020. Severe cases were defined as tachypnea (≥30 breaths/min), oxygen saturation at rest ≤93%, or PaO2/FIO2 < 300 mm Hg; critical cases (ICU) were defined as respiratory failure requiring mechanical ventilation, shock, or nonrespiratory organ failure. Bedside ultrasound examine was performed upon order by the physicians. An ultrasound scanner equipped with an L12-5/S5-1 probe (EPIQ 7C, Philips Medical Systems, Andover, MA, USA,) or a Mindray portable Ultrasound (M9, GD, China, equipped with an L10-3 probe) had been used to examine bilateral common femoral, deep, and superficial femoral, and the popliteal veins as well as the posterior tib- ial, peroneal, and calf muscle veins. Clinic features, laboratory findings, and outcomes of the patients were collected during hospitalization and analyzed by two independent analysts.
Altogether 235 COVID-19 patients were recruited in this study, including 131 (55.7%) men and 104 (44.3%) women, and the mean age was 67.0 years (SD 11.3). Of them, 154 (65.5%) patients were in severe/critical condition and 71 (30.2%) were dead. The mean time from symptom onset to hospital admission was 12.7 ± 7.8 days. Bedside ultrasound has been used to performed lower limb vascular examination for 204 patients, and DVT had been observed in 104 (50.1%) of them. As shown in Table  S1, we observed no significant difference in gender, body mass index (BMI), or days from symptom onset to hospital admission between the DVT group and non-DVT group. In contrast, patients with DVT were associated with older ages (67.0 ± 11.3 vs. 59.7 ± 15.5, p < 0.001) and longer bedridden time ( , p = 0.002) and low molecular weight heparin, both before (p = 0.002) and after ultrasound diagnosis (p < 0.001). We then used univariate and multivariate logistic regression models to identify the risk factors associated with DVT. As shown in Table 1, univariate analysis revealed that risk factors associated with DVT included age, disease severity, death, bedridden time, diastolic blood pressure (DBP), the numbers of platelet, white blood cells, neutrophil, and lymphocytes, as well as the levels of Ddimer, RT .260]; p = 0.042) were independent risk factors associated with DVT. Using a receiver-operating characteristic analysis, a combination of age, D-dimer, and CK-MB yielded a sensitivity of 79.81% and a specificity of 73.20% for prediction of DVT ( Figure S1).
Together, our results revealed that the viral infection was associated with a transient increased risk of venous thromboembolic events. DVT had been observed in half of the patients, particularly for those who were critically ill, in line with previous studies showing an overall rate of 69% in a French cohort. 4 The actual DVT rate could be even higher as the bedside ultrasound examines had only been performed for a few times, highlighting the important role of DVT in COVID-19 pathogenesis. DVT might lead to more severe thromboembolic diseases such as PE, resulting in increased mortality. Moreover, patients with a history of acute ischemic stroke (AIS) and/or its risk factors are particularly at-risk. In line with this notion, AIS had been observed in 5.7% of the severe cases, which can be immediately linked to increased mortality and poor outcomes, indicating the need of active anticoagulant therapy. Although coagulation test screening (i.e., measurement of D-dimer and fibrinogen levels) is indicative of increased risk of DVT, current guideline do not suggest the use of full-intensity anticoagulation doses unless otherwise clinically indicated. In addition to DVT, myocardial injury is a common phenomenon in patients with COVID-19 as indicated by elevated levels of CK-MB. Our findings that elder age, high levels of D-dimer, and CK-MB are independently associated with DVT, particularly in individuals with preexisting cardiovascular diseases, providing implications for choice, dosing, and laboratory monitoring of antithrombotic therapies. Our results suggested that early intervention and thromboembolic prophylaxis should be considered for those patients.

A C K N O W L E D G M E N T S
This work was supported by the National Natural Science Foundation of China (grant 81922033 to Dr. Zhang; grant 81727805 to Dr. Xie).