Xây dựng mô hình tiên lượng rối loạn đông máu nặng sau ghép gan

Abstract

Objective: To develop a prediction model for severe coagulopathy (SC) in patients after living donor liver transplantation at the 108th Central Military Hospital. Subjects and Methods: A descriptive, retrospective study on 202 patients who underwent liver transplantation from living donors between January 2020 and January 2025. Data collected included clinical and laboratory parameters before, during, and after the liver transplantation surgery. Multivariate logistic regression was used to identify independent prognostic factors after liver transplantation and to build the prediction model. Results: The incidence of severe coagulopathy after liver transplantation was 53.5%. Independent prognostic factors for severe coagulopathy included the pre-transplant Child-Pugh score (OR = 1.48), MELD score (OR = 1.09), pre-transplant platelet count (OR=0.83), and blood loss during surgery (OR = 1.04). The prediction model showed high accuracy with an AUC of 0.944 and a MAE of 0.014 when using the Bootstrap method with 1000 iterations. Conclusion: The prediction model, built from independent prognostic factors, demonstrates good predictive value and potential clinical application to identify high-risk patients for severe coagulopathy after liver transplantation. It can help improve resuscitation strategies and minimize postoperative complications.