Abstract
BACKGROUND: Deep venous thrombosis (DVT), known to be a major factor in poor outcomes and death rates, is common after polytrauma with traumatic brain injury (TBI). In this study, a nomogram will be developed to predict the risk of DVT in polytrauma patients with TBI, since there is currently no specific and convenient diagnostic method. METHODS: A retrospective and observational trial was conducted between November 2021 and May 2023. The predictive model was created using a group of 349 polytrauma patients with TBI in a training set, with data collected between November 2021 and August 2022. A nomogram was presented after using multivariable logistic regression analysis to create the predictive model. Validation of the model was conducted internally. A separate group for validation included 298 patients seen consecutively between August 2022 and May 2023. RESULTS: A total of 647 trauma patients were included in the study. Out of these, 349 individuals were part of the training group, while 298 were part of the validation group. Training cohorts reported 32.1% and validation cohorts reported 31.9% DVT. Age, Smoking, Injury Severity Score (ISS), Glasgow Coma Scale (GCS), D-dimer, Mechanical ventilation (MV) and Application of Vasoactive Drugs (AVD) comprised the individualized prediction nomogram. The model exhibited strong discrimination, achieving a C-index of 0.783 and a statistically insignificant result (P=0.216) following the Hosmer-Lemeshow test. Nomogram calibration plots and decision curve analysis showed the nomogram's utility in predicting DVT. CONCLUSION: Our study characterized the incidence of DVT in polytrauma patients with TBI and further emphasized that it represented a substantial health concern, as evidenced by its frequency. Using this nomogram, it is possible to predict DVT in polytrauma patients with TBI based on demographics and clinical risk factors.