COVID-19 and coagulation: bleeding and thrombotic manifestations of SARS-CoV-2 infection

Abstract

Patients with coronavirus disease 2019 (COVID-19) have elevated D-dimer levels. Early reports describe high venous thromboembolism (VTE) and disseminated intravascular coagulation (DIC) rates, but data are limited. This multicenter retrospective study describes the rate and severity of hemostatic and thrombotic complications of 400 hospital-admitted COVID-19 patients (144 critically ill) primarily receiving standard-dose prophylactic anticoagulation. Coagulation and inflammatory parameters were compared between patients with and without coagulation-associated complications. Multivariable logistic models examined the utility of these markers in predicting coagulation-associated complications, critical illness, and death. The radiographically confirmed VTE rate was 4.8% (95% confidence interval [CI], 2.9-7.3), and the overall thrombotic complication rate was 9.5% (95% CI, 6.8-12.8). The overall and major bleeding rates were 4.8% (95% CI, 2.9-7.3) and 2.3% (95% CI, 1.0-4.2), respectively. In the critically ill, radiographically confirmed VTE and major bleeding rates were 7.6% (95% CI, 3.9-13.3) and 5.6% (95% CI, 2.4-10.7), respectively. Elevated D-dimer at initial presentation was predictive of coagulation-associated complications during hospitalization (D-dimer >2500 ng/mL, adjusted odds ratio [OR] for thrombosis, 6.79 [95% CI, 2.39-19.30]; adjusted OR for bleeding, 3.56 [95% CI, 1.01-12.66]), critical illness, and death. Additional markers at initial presentation predictive of thrombosis during hospitalization included platelet count >450 × 109/L (adjusted OR, 3.56 [95% CI, 1.27-9.97]), C-reactive protein (CRP) >100 mg/L (adjusted OR, 2.71 [95% CI, 1.26-5.86]), and erythrocyte sedimentation rate (ESR) >40 mm/h (adjusted OR, 2.64 [95% CI, 1.07-6.51]). ESR, CRP, fibrinogen, ferritin, and procalcitonin were higher in patients with thrombotic complications than in those without. DIC, clinically relevant thrombocytopenia, and reduced fibrinogen were rare and were associated with significant bleeding manifestations. Given the observed bleeding rates, randomized trials are needed to determine any potential benefit of intensified anticoagulant prophylaxis in COVID-19 patients.

Graphical abstract

Figure 1.

Correlation matrix showing the strength of correlation between peak values of D-dimer and evaluated inflammatory parameters. Values in cells are Spearman correlation coefficients. All correlations were statistically significant (P < .0001). 95% CIs for each correlation coefficient can be found in supplemental Table 6.

Figure 2.

Association of coagulation and inflammatory parameters at initial presentation with thrombotic complications during hospitalization in univariable and multivariable analyses. aPTT, activated PTT.

Figure 3.

Association of coagulation and inflammatory parameters at initial presentation with bleeding events during hospitalization in univariable and multivariable analyses. aPTT, activated PTT.

Figure 4.

Association of coagulation and inflammatory parameters at initial presentation with critical illness in univariable and multivariable analyses. aPTT, activated PTT.

Figure 5.

Association of coagulation and inflammatory parameters at initial presentation with mortality during hospitalization in univariable and multivariable analyses. Analysis was limited to those patients reaching a terminal end point (hospital discharge or death) by the end of the study period (n = 252). aPTT, activated PTT.