It's increasingly clear that critically ill patients with COVID-19 can develop a pro-thrombotic form of DIC which places them at a dramatically increased risk of thrombosis. Thrombotic events may include pulmonary micro-vascular thrombosis (reported in some autopsies) or macro-vascular thrombosis (e.g., deep vein thrombosis). To date, an evidentiary vacuum has left it highly controversial as to when patients should receive therapeutic anticoagulation.
Cui et al: Prevalence of venous thromboembolism in patients with severe novel coronavirus pneumonia
This is a single-center retrospective study describing 81 patients with COVID-19 requiring ICU admission.1 All patients were evaluated for lower extremity deep vein thrombosis using ultrasonography.
Baseline demographic data is sparse (table below). For example, it's unclear how many patients were intubated. Based on the very high discharge rate (85%), it's likely that this was a cohort of patients with relatively mild disease (when compared to other ICU cohorts).
25% of patients developed DVT. DVT correlated with indicators of disease severity and DIC (e.g., older age, lower lymphocyte count, and prolonged APTT). The strongest correlation was with D-dimer:
The authors subsequently tested various D-dimer cutoff values for the ability to predict DVT occurrence. The best cutoff was 1,500 ng/ml (with a sensitivity of 85% and a specificity of 89%).
This study has numerous weaknesses:
- It's unclear how frequently DVT studies were performed. If testing was performed only once, or very infrequently – then this could miss DVTs.
- Patients didn't receive prophylactic anticoagulation to prevent DVT. This likely increased the observed rate of DVT.
- This cohort of patients appears less ill than most ICU cohorts, which could decrease the rate of DVT (when compared to sicker cohorts of patients).
- The occurrence of pulmonary emboli wasn't systematically reported or investigated. It's likely that some patients without DVT did have a PE, thus under-estimating the burden of venous thromboembolic disease.
- The rate of venous thromboembolic disease in patients with severe COVID is substantial.
- D-dimer appears to be the best single predictor of patients who will develop venous thromboembolic disease.
- In this patient cohort, a D-dimer >1,500 ng/ml had an 85% sensitivity and 89% specificity for predicting which patients would develop DVT. This supports the concept of empiric anticoagulation for patients with markedly elevated D-dimers (especially in situations where frequent CT angiography is impossible due to logistic restraints).
- Numerous methodological limitations in this study prevent it from being definitive (e.g., lack of screening for pulmonary emboli may have led to under-diagnosis of venous thromboembolic disease).
- Until additional data is available, when to initiate full anticoagulation will remain controversial. For now, these decisions may be judged on a patient-by-patient basis, considering both risks of thrombosis and hemorrhage. Among patients without risk factors for hemorrhage, empiric anticoagulation may be reasonable for patients with D-dimer levels above ~1,500 ng/ml.
- 1.Cui S, Chen S, Li X, Liu S, Wang F. Prevalence of venous thromboembolism in patients with severe novel coronavirus pneumonia. J Thromb Haemost. April 2020. doi:10.1111/jth.14830