AKI persistence at 48 h predicts mortality in patients with acute on chronic liver failure.

BACKGROUND AND AIM: Management of acute kidney injury (AKI) in cirrhotics has undergone a paradigm change. We evaluated the impact of AKI persistence at 48 h on outcome in patients with acute on chronic liver failure (ACLF).

METHODS: Consecutive patients with ACLF (n = 373) were prospectively followed. AKI was defined as increase in serum creatinine of 0.3 mg/dl or 1.5- to 2-fold from baseline. Persistent AKI was defined as nonresponsive AKI at 48 h with respect to admission serum creatinine.

RESULTS: AKI at admission was present in 177 (47.5 %) patients. At 48 h, 73 % patients had persistent AKI and 27 % had responsive AKI. High Model for End-Stage Liver Disease (MELD) (≥26) [p, odds ratio (OR), 95 % confidence interval (CI)] [<0.001, 3.65 (2.1-3.67)], systemic inflammatory response syndrome (SIRS) [0.03, 1.6 (1.02-21.6)], and age (≥42 years) [0.03, 1.84 (1.19-2.85)] were significant predictors of AKI persistence. Persistent AKI was associated with significantly higher in-hospital mortality [p < 0.001, hazard ratio (HR) 1.7, 95 % CI 1.32-2.27]. We further found a lower cutoff for serum creatinine of 1.14 mg/dl at 48 h with better sensitivity of 61 %, specificity of 61 %, and likelihood ratio (LR+) of 1.6, correctly classifying 61 %, as against the conventional cutoff of 1.5 mg/dl with sensitivity of 37 %, specificity of 57 %, and LR+ of 3.3, correctly classifying 56 %. This new cutoff also predicted mortality with higher odds (OR 2.4, 95 % CI 1.3-4.8) as compared with the conventional cutoff (OR 2.1, 95 % CI 1.1-4.1).

CONCLUSION: AKI persistence at 48 h predicts mortality better than serum creatinine of 1.5 mg/dl in patients with ACLF. Serum creatinine value of 1.14 mg/dl and smaller increases in its value should be considered for risk stratification of patients with ACLF for interventional strategies.