Multidimensional preoperative prediction of the evolutionary risk associated with portal hypertension surgery in cirrhosis

Abstract

Introduction: Portal hypertension surgery leads to major complications and postoperative mortality. It therefore requires certain specifications regarding accuracy of prognostic tools. Purpose: The correlation between classic prognostic tools and the risk of severe clinical evolution associated with portal hypertension surgery. Material and methods: The study included 98 splenectomised patients with liver cirrhosis. It was estimated 2 classic scores due to their frequent clinical application, Child and MELD scores and a predictive model of severe evolution risk- BEA (BASELINE Event Anticipation score). Each of them was correlated with postoperative results. Results: Prevalence of scores at enrollment was: Child A/B/C=9/86/5%; MELD<9/10-19/20-29= 71/23/6%; BEA-A/B/C =31/59/10%: mortality 3,1%, 21 complications. We noticed a direct correlation between score classes/complicated evolution rate of the disease and deaths. We documented a minimal superiority of Child score vs MELD score and a lower performance of BEA score(0,91, 0,88 and 0,76) with an underestimation of evolutive risk. The evolution and mortality after a month since surgery were the best predicted by MELD score and after 12 months by the Child score. We remarked: for 21% of the cases the cirrhosis evolution was uncoded by the score, for 56%- converted to a superior class. Conclusions: The prognosis evaluation of an assisted surgical cirrhotic patient encounters a lot of difficulties. Further studies must be conducted for a potential improvement in the results.