Predictive Analysis of Cardiometabolic Risks in Liver Transplantation — A Case Study in Uruguay

Abstract

Cardiovascular diseases are the leading cause of mortality worldwide. In 2021, an estimated 48 million individuals in Latin America were living with heart and circulatory diseases. In the context of liver transplantation, cardiometabolic risk factors play a crucial role not only during the procedure but also in the long-term post-transplantation period, significantly impacting patient survival and recovery. This study analyzes a cohort from the National Liver Transplantation Program of Uruguay, employing machine learning to predict the occurrence of post-transplant cardiometabolic diseases based on pre-transplant health indicators. Over a five-year period, multiple machine learning models were evaluated, with the Extra Trees algorithm achieving the highest predictive accuracy of 88% (AUC: 0.94). The findings highlight the potential of predictive analytics in improving early risk assessment and preventive strategies, ultimately enhancing the prediction of patient outcomes in liver transplantation.Clinical Relevance— This is the first national-level study validating machine learning algorithms for cardiometabolic risk prediction in liver transplantation patients within the National Liver Transplantation Program in Uruguay. By leveraging pretransplant clinical data, the proposed model provides a data-driven approach for early risk stratification, supporting clinicians in making informed decisions to mitigate post-transplant cardiometabolic complications.