P04 Admission serum lactate is a strong predictor of outcome in cirrhotics admitted to intensive care unit, and when added to the liver-specific scores of model for end-stage liver disease or UK model for end-stage liver disease, improves their respective predictive value

Introduction Accurate prognostic indicators of patient survival in an intensive care unit (ICU) help guide clinical decision-making. Factors known to portend poor prognosis in acutely ill cirrhotics in ICU include the need for mechanical ventilation, development of shock, renal failure and sequential increase in the number of failing organs. While serum lactate is now an established marker of survival and/or the need for transplantation in fulminant liver failure, its impact on critically ill cirrhotics is less well known. Method We retrospectively studied 133 consecutive acutely ill cirrhotics admitted to the ICU between 2005 and 2008 at the Royal Free Hospital, a tertiary referral centre in liver diseases and transplantation. Data were collected on demographic variables, aetiology of liver disease, liver-specific prognostic scores (Child–Turcotte–Pugh (CTP), model for end-stage liver disease (MELD), UK model for end-stage liver disease (UKELD)), and acute illness scores (acute physiological score and chronic health point (APACHE II), sequential organ failure assessment score (SOFA)). In addition, serum lactate levels at 0, 24 and 48 h were also recorded. Multivariable logistic regression analysis was performed, and the discrimination ability of each of the above-mentioned scoring models in predicting ICU and hospital survival of these patients was evaluated using the area under the receiver operating characteristic (ROC) curve. Results The ICU and hospital non-survivors—43/133 (32.3%) and 57/133 (43.4%) respectively—had similar demographic features as the survivors, but had significantly higher mean admission MELD, UKELD, SOFA and APACHE II scores, as well serum lactate levels on admission. Serum lactate at admission and particularly at 24 h had a better discriminative accuracy for mortality (AUC=0.737 and 0.764) compared with liver-specific prognostic scores, MELD (AUC=0.732 and 0.720), MELD-Na (AUC=0.338 and 0.554) and UKELD (AUC=0.698 and 0.695). Acute illness scores exhibited a rather poorer predictive power, both APACHE II (AUC=0.632 and 0.571) and SOFA (AUC=0.688 and 0.716). Adding lactate to MELD and UKELD scores further improved their outcome prediction potential (AUC MELD-lactate=0.737 and UKELD-lactate=0.717). Conclusion Serum lactate is a powerful independent tool in predicting survival of acutely ill cirrhotics on ICU. Persistent hyperlactataemia after aggressive resuscitation for 24 h may reflect native liver9s inability to metabolise it. In that case, should lactate not be incorporated in the liver function scoring models such as CTP, MELD or UKELD?