Post-operative acute renal failure is a severe post-operative complication and is associated with high mortality. The enhanced prediction score, including pre-as well as intra-operative predictors accurately predicted ARF following hepatic surgery. This prediction score allows early identification of patients at high risk of ARF and may support decision-making for protective kidney treatment.

To enhance and validate an already pre-existing score accurately predicting post-operative acute renal failure (ARF) after hepatic surgery

We will enhance a pre-existing score predicting ARF based on pre-operative as well as intra-operative predictors.

Development process: we will identify the strongest predictors of ARF in a multivariable logistic regression model followed by a stepwise backward logistic regression analysis and bootstrapping.

Validation process: we will perform an internal validation by calibrating the prediction model as well as by k-fold cross validation (c statistics) and bootstrapping. Additionally, we will calculate the discrimination by the area under the curve (AUC).

Decision curve analysis: Furthermore we will perform a decision curve analysis to evaluate the clinical consequences of both prediction scores whether a patient with increased ARF risk would post-operative benefit of a treatment on the ICU.

Inclusion Criteria:

• > 18 years
• scheduled for liver surgery
• benign as well as malignant diseases

Exclusion Criteria:

• liver trauma
• incomplete data sets
• pre-operative chronic renal failure requiring hemodialysis

We included consecutively patients undergoing any type of hepatic surgery for benign as well as malignant diseases between July first, 2002 and October 31, 2007 in a single tertiary care center (Swiss Hepato-Pancreato-Biliary (HPB) Centre, University Hospital of Zurich, Switzerland.