Lactate and number of organ failures predict intensive care unit mortality in patients with acute-on-chronic liver failure

Background and Aims Patients with acute-on-chronic liver failure (ACLF) have high mortality rates. Most prognostic scores were not developed for the intensive care unit (ICU) setting. We aimed to improve risk stratification for patients with ACLF in the ICU. Methods A training set with 240 patients with cirrhosis and organ failures (Chronic Liver Failure Sequential Organ Failure Assessment score [CLIF-SOFA]) from Curry Cabral Hospital (Portugal) and University of Alberta Hospital (Canada) in 2010-2016 was used to derive a prognostic model for ICU mortality. A validation set with 237 patients with cirrhosis and organ failures from Vancouver General Hospital (Canada) in 2000-2011 was used to evaluate its performance. Results Amongst patients in the training set, ICU and hospital mortality rates were 39.2% and 54.6% respectively. Median lactate (4.4 vs 2.5 mmol/L) and number of organ failures (3 vs 2) on admission to ICU were associated with higher likelihood of ICU mortality (P < 0.001 for both). The lactate and organ failures predictive model (LacOF) was derived to predict ICU mortality: -2.420 + 0.072 x lactate + 0.569 x number of organ failures (area under-the-curve [AUC], 0.76). In the validation set, the LacOF model discriminative ability (AUC, 0.85) outperformed the CLIF-SOFA (AUC, 0.79), Chronic Liver Failure Consortium Acute-on-Chronic Liver Failure (AUC, 0.73), Model for End-stage Liver Disease score (AUC, 0.78) and Acute Physiology and Chronic Health Evaluation II scores (AUC, 0.74; P < 0.05 for all). The LacOF model calibration was good up to the 25% likelihood of ICU mortality. Conclusions In patients with ACLF, lactate and number of organ failures on admission to ICU are useful to predict ICU mortality. This early prognostic evaluation may help to better stratify the risk of ICU mortality and thus optimize organ support strategies.