[Predictive value of von Willebrand factor for venous thromboembolism in critically ill patients based on propensity score matching].

OBJECTIVE

To analyze the predictive value of von Willebrand factor (vWF) for venous thromboembolism (VTE) of patients in intensive care unit (ICU) by using propensity score matching (PSM).

METHODS

Patients admitted to ICU of the Second Affiliated Hospital of Kunming Medical University from December 2020 to June 2022 who stayed in ICU for ≥72 hours and underwent daily bedside vascular ultrasound screening were included. Baseline data such as age, gender, primary disease, and chronic comorbidities were collected. Coagulation indexes before admission to ICU and 24 hours and 48 hours after ICU admission were collected, including prothrombin time (PT), activated partial thromboplastin time (APTT), thrombin time (TT), international normalized ratio (INR), fibrinogen (Fib), fibrin monomer (FM), vWF, D-dimer, antithrombin III (ATIII), etc. Patients were divided into VTE group and non-VTE group according to whether they had VTE or not [diagnosis of VTE: patients underwent daily ultrasound screening of bedside blood vessels (both upper and lower limbs, visceral veins), and those suspected of having thrombosis were confirmed by ultrasonographer or pulmonary angiography]. Using PSM analysis method, the VTE group was used as the benchmark to conduct 1 : 1 matching of age, whether there was malignant tumor, whether there was infection, whether there was diabetes, and coagulation indicators before admission to ICU. Finally, the cases with balanced covariates between the two groups were obtained. The risk factors of VTE were analyzed by multivariate Logistic regression analysis. Receiver operator characteristic curve (ROC curve) was drawn to evaluate the predictive value of vWF in the occurrence of VTE in critically ill patients.

RESULTS

A total of 120 patients were enrolled, of which 18 (15.0%) were diagnosed with VTE within 72 hours after admission to ICU, and 102 (85.0%) were not found to have thrombus in ICU. Before PSM, there were significant differences in age, gender, proportion of malignant tumor and infection, and coagulation indexes between VTE group and non-VTE group. After PSM, 14 pairs were successfully matched, and the unbalanced covariables between the two groups reached equilibrium. Multivariate Logistic regression analysis showed that vWF was an independent risk factor for VTE at 48 hours after ICU admission in critically ill patients [odds ratio (OR) = 1.165, 95% confidence interval (95%CI) was 1.000-1.025, P = 0.004]. ROC curve analysis showed that the area under the ROC curve (AUC) of vWF at 48 hours after ICU admission for predicting VTE was 0.782, 95%CI was 0.618-0.945, P = 0.007. When the optimal cut-off value was 312.12%, the sensitivity was 67.7% and the specificity was 93.0.

CONCLUSIONS

Dynamic monitoring of vWF is helpful to predict the occurrence of VTE in ICU patients, and vWF at 48 hours after ICU admission has certain value in predicting the occurrence of VTE.