Predictive Model for Deep Vein Thrombosis Caused by Closed Lower Limb Fracture After Thromboprophylactic Treatment-WITHDRAWN

Purpose: Currently, there are still no convincing clinical models predicting closed lower extremity fracture-associated deep vein thrombosis in patients treated through thromboprophylactic methods. We aimed at using 2 retrospective cohorts to develop and externally verify a clinical prediction model for deep vein thrombosis in patients treated with anticoagulants after suffering closed lower extremity fractures.

Methods: We evaluated the patients pre- and postoperatively, to accurately determine the predictive power of the biomarkers and clinical risk factors. Two retrospective cohorts were used for the development and external verification of a pre-operative clinical prediction model (development: n = 2253; verification: n = 833) and postoperative clinical prediction model (development: n = 1422; verification: n = 449), respectively.

Results: The C-indices were used to show the predicted incidence of objective thrombosis at the pre- and postoperative stages, which were then compared with the observed incidence of thrombosis in both cohorts. Biomarkers and clinical indicators were included in pre- and postoperative nomograms, which were adequately calibrated in both cohorts. The cross-validated C-indices of the pre- and postoperative clinical prediction models in the verification cohort were 0.706 (95% confidence interval [Cl], 0.67-0.74) and 0.875 (95% Cl, 0.84-0.91), respectively.

Conclusion: We present our findings of novel pre- and postoperative nomograms for the prediction of deep vein thrombosis in patients who received thromboprophylaxis after suffering closed lower extremity fractures.

Graphic