Acute renal failure in cirrhosis: Is it as bad as we think?

Abstract

Introduction: Acute renal failure (ARF) is associated with a mortality of 50–60% in critically ill patients admitted to the intensive care unit (ICU). Prerenal causes and acute tubular necrosis (ATN) account for more than 85% of cases and are potentially reversible. ARF frequently complicates cirrhosis, is often attributed to hepatorenal syndrome (HRS), which may preclude aggressive treatment with its mortality up to 90%. Aims & Methods: The aim of this study was to identify factors that may predispose to or precipitate ARF in cirrhosis, and determine outcome and mortality. A retrospective review of cirrhotic patients admitted with or developing renal impairment (defined as serum creatinine >130 μmol/l or oliguria <500 ml/24 h) from October 1999–April 2004. Patients with bleeding gastro-oesophageal varices were excluded. Demographic details, cause of ARF, potential early warning features, management and outcome were recorded. Results: Eighty patients, median age 52 years (25–84), 46 male, median MELD 26 (7–43). Alcohol was causal in 88.8%. ARF occurred in 41/80 (51.3%) on admission, or a median of 6 days after admission (1–34). Median serum creatinine at onset of renal impairment 172 μmol/l (60–589). An identifiable precipitant was found in 62/80 (77.5%) and were concurrent in 51%; nephrotoxic drugs 50%, sepsis 45% (culture positive 27/36), recent paracentesis (preceding month) 28.8%, fluid loss 26.3%, and spontaneous bacterial peritonitis 13.3%. HRS occurred in 17.5%. No parenchymal renal disease or obstructive uropathy was seen. 54% were hyponatraemic (serum sodium <130 μmol/l) at onset of ARF. ARF was heralded by a fall in median mean arterial pressure (MAP) of 14.1 mm Hg from admission (p<0.001, CI 6.8–19.7). 71.2% received volume expansion, 91.1% terlipressin, 79.7% salt poor albumin, 93.4% antibiotics, 7.5% MARS/renal support. MAP did not rise significantly 24 h after initiation of treatment (p = 0.56). 28 day mortality was 61%. HRS patients had higher MELD scores than non-HRS patients (30 v 25; p = 0.0152) but 28 day mortality was not significantly different (64.3% v 60.6% respectively, p = 0.797). Conclusion: ARF in cirrhosis is rarely due to HRS but is most commonly pre-renal in origin and multifactorial. Potential early warning features include hyponatraemia and a falling MAP. Current treatment regimes may not be aggressive enough to reverse renal hypoperfusion. Despite this mortality for our cirrhotic cohort was similar to non-cirrhotic patients admitted to ITU with ARF. Development of ARF in a cirrhotic patient should not preclude aggressive treatment.