Robust Predictive Performance of the SALT-M Score for Clinical Outcomes in Asian Patients With Acute-on-Chronic Liver Failure.

Abstract

Acute-on-chronic liver failure (ACLF) is a syndrome of patients with chronic liver disease presenting with multiple organ failures. Recently, Sundaram-ACLF-LT Mortality (SALT-M) score has been developed to predict 1-year post-liver transplantation mortality. We validated the SALT-M score in a large-volume, Asian single-centre cohort. We validated the SALT-M score in a large-volume, Asian single-centre cohort. We analysed 224 patients of ACLF grade 2-3. Area under the receiver operating characteristic curve (AUROC) and concordance index (c-index) were used to assess and compare the predictability of posttransplant mortality of SALT-M and other scores. Moreover, we compared the survivals of patients with high and low SALT-M, in conjunction with MELD score and ACLF grade. The AUROC for prediction of 1-year post-LT survival was higher in SALT-M (0.691) than in MELD, MELD-Na, MELD 3.0 and delta-MELD. Similarly, the c-index of the SALT-M (0.650) was higher than aforementioned MELD systems. When categorised by the cut-off of SALT-M ≥ 20 and MELD ≥ 30, patients with high SALT-M exhibited lower post-LT survival than those with low SALT-M scores regardless of high or low MELD (40.0% for high SALT-M/high MELD vs. 42.9% for high SALT-M/low MELD vs. 73.8% for low SALT-M/high MELD vs. 63.7% for low SALT-M/low MELD, p < 0.001). In patients with ACLF grade 3, SALT-M effectively stratified the posttransplant mortality (39.4% for high SALT-M vs. 63.1% for low SALT-M, p = 0.018). SALT-M outperformed previous MELD systems for predicting posttransplant mortality in Asian LT cohort with severe ACLF. Transplantability for patients with severe ACLF could be determined based on SALT-M.