Liver dysfunction by model for end-stage liver disease score improves mortality prediction in injured patients with cirrhosis.

BACKGROUND

Cirrhosis is associated with poor outcomes in the trauma setting. We aimed to evaluate the utility of Model for End-Stage Liver Disease (MELD) in assessing additional mortality risk in trauma patients with cirrhosis.

METHODS

Injured patients with liver dysfunction were identified by hospital and trauma registry query. Presence of cirrhosis was confirmed by laparotomy, biopsy, or imaging. MELD classification, Child-Turcotte-Pugh (CTP) classification, Injury Severity Score (ISS), and Trauma ISS (TRISS) were recorded, and the primary outcome variable was hospital mortality. We assessed the validity of the four scoring systems in prediction of mortality, individually and in combinations, by comparing the areas under receiver operating characteristic curves (AUC), which is the probability, for scores that increase with the risk of death that a randomly chosen deceased subject will score higher than a randomly chosen living subject.

RESULTS

A total of 163 patients with confirmed cirrhosis were included. ISS (AUC = 0.849, p < 0.001) and TRISS (AUC = 0.826, p < 0.001) were the strongest predictors of mortality. MELD (AUC = 0.725) was not a significantly stronger predictor of mortality than CTP (AUC = 0.639; p = 0.38). ISS + MELD (AUC = 0.891) and ISS + CTP (AUC = 0.897) were stronger predictors than ISS alone (AUC = 0.849; p < 0.001) for both. The MELD score was more available from the records than the CTP score (91.4% vs. 75.5%).

CONCLUSION

In trauma patients with cirrhosis, a score that evaluates the degree of liver dysfunction enhances the ability of ISS alone to predict mortality. The MELD score is more readily available than the CTP score for the prediction of mortality in trauma patients.