PB0635 - A Nomogram Predictive Model for Prediction of Venous thrombo embolism in Inpatients: A Retrospective Study

Background: Venous thrombo embolism (VTE) and pulmonary embolus are associated with hospital-related morbidity and mortality, and are recognized as complications in patients with traumatic injury.

Aims: The present study aimed to develop and validate a nomogram predictive model based on thrombus-related markers to predict VTE in inpatients.

Methods: We enrolled 270 patients under deep vein ultrasonography from January 2021 to October 2021, including a training cohort (60/130 with/without VTE) and a validation cohort (25/55 with/without VTE). The least absolute shrinkage and selection operator (LASSO) regression model was applied to optimize model feature selection to predict VTE. Multivariable logistic regression analysis was applied to construct the final VTE predictive model which was visualized with nomogram. The performance and clinical utility of the nomogram were evaluated using the area under the curve (AUC) from a receiver operating characteristic (ROC) curve analysis, calibration curve, and decision curve analysis (DCA).

Results: Based on LASSO regression and multivariable logistic regression analysis, prothrombin time (PT), activated partial thromboplastin time (APTT), factor VIII (FVIII), Alanine Transaminase (ALT), Gamma Glutamyl Transferase (GGT) and homocysteine (HCY) were finally included in the VTE predictive model and visualized using nomogram. The Hosmer-Lemeshow test yielded a nonsignificant P value of 0.1777, suggested that the model was of goodness-of-fit between the predicted and actually values. The predictive model achieved an AUC of 0.946 (95% CI, 0.913-0.978) and 0.916 (95% CI, 0.857-0.975) for the training and validation cohorts, compared to 0.718 (95% CI, 0.64-0.797) and 0.806 (95% CI, 0.703-0.909) by the Padua scale score. In addition, the DCA indicated that the predictive model exhibited greater net benefits for the identification of VTE than the Padua scale score.

Conclusion(s): Compared with the Padua scale score, the novel nomogram displayed a superior performance in terms of predictive accuracy, discrimination capability and clinical utility in predicting VTE.