Early liver parameter increase after transjugular intrahepatic portosystemic shunt (TIPS): hepatic vascular risk factors and development of an exploratory prediction model.

PURPOSE: To investigate early laboratory changes following Transjugular Intrahepatic Portosystemic Shunt (TIPS) and their associations with hepatic vessel diameters. Additionally, to develop a model to predict patients at high-risk for excessive liver parameter increase. MATERIALS AND METHODS: A total of 256 TIPS procedures performed between 2017 and 2024 were retrospectively enrolled and divided into a development (n = 208) and temporally separated internal validation cohort (n = 48). Diameters of portal vein and hepatic arteries were measured in pre-procedural computed tomography scans. Laboratory parameters (INR, ALAT, ASAT, GGT, and bilirubin) were assessed pre-procedurally, on post-procedural days 1-2, and between days 4-7. Patients with ≥ 2 parameters exceeding the 85th percentile of percentage change on days 1-2 were classified as high-risk. A prediction model for the identification of high-risk patients was developed and validated. RESULTS: All laboratory parameters significantly increased post-procedurally (p < 0.001), with 41 patients classified as high-risk. ALAT and ASAT showed the highest increase (up to +392 +/- 1118% and +474 +/- 1125%, respectively). Hepatic artery and portal vein diameters were predominantly negatively correlated with laboratory changes (p < 0.05). The prediction model identified common hepatic artery diameter, portal vein thrombosis, and baseline bilirubin as independent high-risk predictors, achieving an AUC of 0.782 and 0.748 in the development and internal validation cohort, respectively. CONCLUSION: TIPS significantly impacts liver metabolism, with marked early alterations in liver parameters that inversely correlate with hepatic vessel diameters. A prediction model based on common hepatic artery diameter, portal vein thrombosis status, and baseline bilirubin may support identification of high-risk patients and potentially help optimize post-procedural management.