Role of serum creatinine and prognostic scoring systems in assessing hospital mortality in critically ill cirrhotic patients with upper gastrointestinal bleeding.

BACKGROUND End-stage liver disease is frequently complicated by episodes of gastrointestinal hemorrhage that are often associated with multiple organ dysfunction and require intensive care. This study aimed to identify specific predictors of hospital mortality in critically ill cirrhotic patients with gastrointestinal bleeding, and compare the prediction accuracy of the Child-Pugh score and two illness severity scoring systems frequently used for intensive care unit (ICU) patients. METHODS 76 patients with liver cirrhosis and upper gastrointestinal bleeding were admitted to the ICU from April 2001 to March 2002. In addition, 27 demographic, clinical and laboratory variables, including parameters assessing liver and renal function and systemic hemodynamics, were analyzed as survival predicators. Finally, information required, calculating the Child-Pugh, Sequential Organ Failure Assessment (SOFA), and Acute Physiology and Chronic Health Evaluation (APACHE) III score on the 1st day of ICU admission, was gathered prospectively. RESULTS Overall, hospital mortality was 68.4%. Liver disease was generally attributed to hepatitis B viral infection. Furthermore, multiple logistic regression analysis showed that mean arterial pressure (MAP), Child-Pugh points, and serum creatinine (Cr) were significantly related to prognosis. The SOFA and APACHE III models displayed good areas under the receiver operating characteristic (ROC) curve. CONCLUSION The rise of serum Cr levels above 1.5 mg/dL is common, and indicates a poor prognosis for critically ill cirrhotic patients with gastrointestinal bleeding. SOFA is a straightforward approach with excellent prognostic abilities for this homogeneous patient subset.