Background/objectives: to compare the prognostic accuracy for 28 and 90-day transplant-free mortality of a modified CLIF-SOFA score (including a dynamic definition of acute kidney injury) with that of the classic CLIF-SOFA score and KDIGO score for acute kidney injury in patients with acute decompensation of cirrhosis.
Methods: A retrospective analysis of all admissions of acutely decompensated patients with cirrhosis was carried out from January 2012 to December 2014. Classic and modified CLIF-SOFA scores were analyzed, as well as acute kidney injury diagnosis using the KDIGO score regarding their accuracy for 28- and 90-day transplant free mortality prediction.
Results: 108 admissions were analyzed. Acute kidney injury diagnosis was met in 37 (34%) patients. Acute-on-chronic liver failure was diagnosed in 59 (55%) patients using the classic CLIF-SOFA score; and in 64 (59%) patients using the modified CLIF-SOFA score. Both CLIF-SOFA scores were highly effective in predicting 28-day transplant-free mortality (AUCROC 0.93 and 0.92, p = 0.34) as well as 90-day transplant-free mortality (AUCROC 0.79 and 0.78, p = 0.78). Acute kidney injury diagnosis had significantly lower accuracy in mortality assessment (28 and 90-day transplant free mortality AUCROC 0.67 [p = 0.002] and 0.63 [p = 0.02]).
Conclusions: To our knowledge, this is the first evidence of the limited impact of modifying the fixed kidney injury definition currently used for acute-on-chronic liver failure.
Keywords: AUCROC, area under the curve of the receiving operating characteristic; CLIF-SOFA, Chronic Liver Failure-Sequential Organ Failure Assessment; HIV, human immunodeficiency virus; INR, international normalized radio; KDIGO, Kidney Disease Improving Global Outcome; MELD, model for end stage liver disease; acute kidney injury; acute-on-chronic liver failure; mortality; prognostic accuracy.